Merge branch 'dygraph' of https://github.com/PaddlePaddle/PaddleOCR into py_inference_doc

ppocr/postprocess/db_postprocess.py

@@ -138,9 +138,9 @@ class DBPostProcess(object):
 
         boxes_batch = []
         for batch_index in range(pred.shape[0]):
-            height, width = shape_list[batch_index]
+            src_h, src_w, ratio_h, ratio_w = shape_list[batch_index]
             boxes, scores = self.boxes_from_bitmap(
-                pred[batch_index], segmentation[batch_index], width, height)
+                pred[batch_index], segmentation[batch_index], src_w, src_h)
 
             boxes_batch.append({'points': boxes})
         return boxes_batch
\ No newline at end of file

ppocr/postprocess/east_postprocess.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import numpy as np
+from .locality_aware_nms import nms_locality
+import cv2
+
+import os
+import sys
+# __dir__ = os.path.dirname(os.path.abspath(__file__))
+# sys.path.append(__dir__)
+# sys.path.append(os.path.abspath(os.path.join(__dir__, '..')))
+
+
+class EASTPostProcess(object):
+    """
+    The post process for EAST.
+    """
+    def __init__(self,
+                 score_thresh=0.8,
+                 cover_thresh=0.1,
+                 nms_thresh=0.2,
+                 **kwargs):
+
+        self.score_thresh = score_thresh
+        self.cover_thresh = cover_thresh
+        self.nms_thresh = nms_thresh
+        
+        # c++ la-nms is faster, but only support python 3.5
+        self.is_python35 = False
+        if sys.version_info.major == 3 and sys.version_info.minor == 5:
+            self.is_python35 = True
+
+    def restore_rectangle_quad(self, origin, geometry):
+        """
+        Restore rectangle from quadrangle.
+        """
+        # quad
+        origin_concat = np.concatenate(
+            (origin, origin, origin, origin), axis=1)  # (n, 8)
+        pred_quads = origin_concat - geometry
+        pred_quads = pred_quads.reshape((-1, 4, 2))  # (n, 4, 2)
+        return pred_quads
+
+    def detect(self,
+               score_map,
+               geo_map,
+               score_thresh=0.8,
+               cover_thresh=0.1,
+               nms_thresh=0.2):
+        """
+        restore text boxes from score map and geo map
+        """
+        score_map = score_map[0]
+        geo_map = np.swapaxes(geo_map, 1, 0)
+        geo_map = np.swapaxes(geo_map, 1, 2)
+        # filter the score map
+        xy_text = np.argwhere(score_map > score_thresh)
+        if len(xy_text) == 0:
+            return []
+        # sort the text boxes via the y axis
+        xy_text = xy_text[np.argsort(xy_text[:, 0])]
+        #restore quad proposals
+        text_box_restored = self.restore_rectangle_quad(
+            xy_text[:, ::-1] * 4, geo_map[xy_text[:, 0], xy_text[:, 1], :])
+        boxes = np.zeros((text_box_restored.shape[0], 9), dtype=np.float32)
+        boxes[:, :8] = text_box_restored.reshape((-1, 8))
+        boxes[:, 8] = score_map[xy_text[:, 0], xy_text[:, 1]]
+        if self.is_python35:
+            import lanms
+            boxes = lanms.merge_quadrangle_n9(boxes, nms_thresh)
+        else:
+            boxes = nms_locality(boxes.astype(np.float64), nms_thresh)
+        if boxes.shape[0] == 0:
+            return []
+        # Here we filter some low score boxes by the average score map, 
+        #   this is different from the orginal paper.
+        for i, box in enumerate(boxes):
+            mask = np.zeros_like(score_map, dtype=np.uint8)
+            cv2.fillPoly(mask, box[:8].reshape(
+                (-1, 4, 2)).astype(np.int32) // 4, 1)
+            boxes[i, 8] = cv2.mean(score_map, mask)[0]
+        boxes = boxes[boxes[:, 8] > cover_thresh]
+        return boxes
+
+    def sort_poly(self, p):
+        """
+        Sort polygons.
+        """
+        min_axis = np.argmin(np.sum(p, axis=1))
+        p = p[[min_axis, (min_axis + 1) % 4,\
+            (min_axis + 2) % 4, (min_axis + 3) % 4]]
+        if abs(p[0, 0] - p[1, 0]) > abs(p[0, 1] - p[1, 1]):
+            return p
+        else:
+            return p[[0, 3, 2, 1]]
+
+    def __call__(self, outs_dict, shape_list):
+        score_list = outs_dict['f_score']
+        geo_list = outs_dict['f_geo']
+        img_num = len(shape_list)
+        dt_boxes_list = []
+        for ino in range(img_num):
+            score = score_list[ino].numpy()
+            geo = geo_list[ino].numpy()
+            boxes = self.detect(
+                score_map=score,
+                geo_map=geo,
+                score_thresh=self.score_thresh,
+                cover_thresh=self.cover_thresh,
+                nms_thresh=self.nms_thresh)
+            boxes_norm = []
+            if len(boxes) > 0:
+                h, w = score.shape[1:]
+                src_h, src_w, ratio_h, ratio_w = shape_list[ino]
+                boxes = boxes[:, :8].reshape((-1, 4, 2))
+                boxes[:, :, 0] /= ratio_w
+                boxes[:, :, 1] /= ratio_h
+                for i_box, box in enumerate(boxes):
+                    box = self.sort_poly(box.astype(np.int32))
+                    if np.linalg.norm(box[0] - box[1]) < 5 \
+                        or np.linalg.norm(box[3] - box[0]) < 5:
+                        continue
+                    boxes_norm.append(box)
+            dt_boxes_list.append({'points': np.array(boxes_norm)})
+        return dt_boxes_list
\ No newline at end of file

ppocr/postprocess/locality_aware_nms.py

+"""
+Locality aware nms.
+"""
+
+import numpy as np
+from shapely.geometry import Polygon
+
+
+def intersection(g, p):
+    """
+    Intersection.
+    """
+    g = Polygon(g[:8].reshape((4, 2)))
+    p = Polygon(p[:8].reshape((4, 2)))
+    g = g.buffer(0)
+    p = p.buffer(0)
+    if not g.is_valid or not p.is_valid:
+        return 0
+    inter = Polygon(g).intersection(Polygon(p)).area
+    union = g.area + p.area - inter
+    if union == 0:
+        return 0
+    else:
+        return inter / union
+
+
+def intersection_iog(g, p):
+    """
+    Intersection_iog.
+    """
+    g = Polygon(g[:8].reshape((4, 2)))
+    p = Polygon(p[:8].reshape((4, 2)))
+    if not g.is_valid or not p.is_valid:
+        return 0
+    inter = Polygon(g).intersection(Polygon(p)).area
+    #union = g.area + p.area - inter
+    union = p.area
+    if union == 0:
+        print("p_area is very small")
+        return 0
+    else:
+        return inter / union
+
+
+def weighted_merge(g, p):
+    """
+    Weighted merge.
+    """
+    g[:8] = (g[8] * g[:8] + p[8] * p[:8]) / (g[8] + p[8])
+    g[8] = (g[8] + p[8])
+    return g
+
+
+def standard_nms(S, thres):
+    """
+    Standard nms.
+    """
+    order = np.argsort(S[:, 8])[::-1]
+    keep = []
+    while order.size > 0:
+        i = order[0]
+        keep.append(i)
+        ovr = np.array([intersection(S[i], S[t]) for t in order[1:]])
+
+        inds = np.where(ovr <= thres)[0]
+        order = order[inds + 1]
+
+    return S[keep]
+
+
+def standard_nms_inds(S, thres):
+    """
+    Standard nms, retun inds.
+    """
+    order = np.argsort(S[:, 8])[::-1]
+    keep = []
+    while order.size > 0:
+        i = order[0]
+        keep.append(i)
+        ovr = np.array([intersection(S[i], S[t]) for t in order[1:]])
+
+        inds = np.where(ovr <= thres)[0]
+        order = order[inds + 1]
+
+    return keep
+
+
+def nms(S, thres):
+    """
+    nms.
+    """
+    order = np.argsort(S[:, 8])[::-1]
+    keep = []
+    while order.size > 0:
+        i = order[0]
+        keep.append(i)
+        ovr = np.array([intersection(S[i], S[t]) for t in order[1:]])
+
+        inds = np.where(ovr <= thres)[0]
+        order = order[inds + 1]
+
+    return keep
+
+
+def soft_nms(boxes_in, Nt_thres=0.3, threshold=0.8, sigma=0.5, method=2):
+    """
+    soft_nms
+    :para boxes_in, N x 9 (coords + score)
+    :para threshould, eliminate cases min score(0.001)
+    :para Nt_thres, iou_threshi
+    :para sigma, gaussian weght
+    :method, linear or gaussian
+    """
+    boxes = boxes_in.copy()
+    N = boxes.shape[0]
+    if N is None or N < 1:
+        return np.array([])
+    pos, maxpos = 0, 0
+    weight = 0.0
+    inds = np.arange(N)
+    tbox, sbox = boxes[0].copy(), boxes[0].copy()
+    for i in range(N):
+        maxscore = boxes[i, 8]
+        maxpos = i
+        tbox = boxes[i].copy()
+        ti = inds[i]
+        pos = i + 1
+        #get max box
+        while pos < N:
+            if maxscore < boxes[pos, 8]:
+                maxscore = boxes[pos, 8]
+                maxpos = pos
+            pos = pos + 1
+        #add max box as a detection
+        boxes[i, :] = boxes[maxpos, :]
+        inds[i] = inds[maxpos]
+        #swap
+        boxes[maxpos, :] = tbox
+        inds[maxpos] = ti
+        tbox = boxes[i].copy()
+        pos = i + 1
+        #NMS iteration
+        while pos < N:
+            sbox = boxes[pos].copy()
+            ts_iou_val = intersection(tbox, sbox)
+            if ts_iou_val > 0:
+                if method == 1:
+                    if ts_iou_val > Nt_thres:
+                        weight = 1 - ts_iou_val
+                    else:
+                        weight = 1
+                elif method == 2:
+                    weight = np.exp(-1.0 * ts_iou_val**2 / sigma)
+                else:
+                    if ts_iou_val > Nt_thres:
+                        weight = 0
+                    else:
+                        weight = 1
+                boxes[pos, 8] = weight * boxes[pos, 8]
+                #if box score falls below thresold, discard the box by
+                #swaping last box update N
+                if boxes[pos, 8] < threshold:
+                    boxes[pos, :] = boxes[N - 1, :]
+                    inds[pos] = inds[N - 1]
+                    N = N - 1
+                    pos = pos - 1
+            pos = pos + 1
+
+    return boxes[:N]
+
+
+def nms_locality(polys, thres=0.3):
+    """
+    locality aware nms of EAST
+    :param polys: a N*9 numpy array. first 8 coordinates, then prob
+    :return: boxes after nms
+    """
+    S = []
+    p = None
+    for g in polys:
+        if p is not None and intersection(g, p) > thres:
+            p = weighted_merge(g, p)
+        else:
+            if p is not None:
+                S.append(p)
+            p = g
+    if p is not None:
+        S.append(p)
+
+    if len(S) == 0:
+        return np.array([])
+    return standard_nms(np.array(S), thres)
+
+
+if __name__ == '__main__':
+    # 343,350,448,135,474,143,369,359
+    print(
+        Polygon(np.array([[343, 350], [448, 135], [474, 143], [369, 359]]))
+        .area)
\ No newline at end of file

ppocr/postprocess/rec_postprocess.py

@@ -23,14 +23,16 @@ class BaseRecLabelDecode(object):
                  character_dict_path=None,
                  character_type='ch',
                  use_space_char=False):
-        support_character_type = ['ch', 'en', 'en_sensitive']
+        support_character_type = [
+            'ch', 'en', 'en_sensitive', 'french', 'german', 'japan', 'korean'
+        ]
         assert character_type in support_character_type, "Only {} are supported now but get {}".format(
-            support_character_type, self.character_str)
+            support_character_type, character_type)
 
         if character_type == "en":
             self.character_str = "0123456789abcdefghijklmnopqrstuvwxyz"
             dict_character = list(self.character_str)
-        elif character_type == "ch":
+        elif character_type in ["ch", "french", "german", "japan", "korean"]:
             self.character_str = ""
             assert character_dict_path is not None, "character_dict_path should not be None when character_type is ch"
             with open(character_dict_path, "rb") as fin:
@@ -150,4 +152,4 @@ class AttnLabelDecode(BaseRecLabelDecode):
         else:
             assert False, "unsupport type %s in get_beg_end_flag_idx" \
                           % beg_or_end
-        return idx
+        return idx
\ No newline at end of file

ppocr/postprocess/sast_postprocess.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import os
+import sys
+__dir__ = os.path.dirname(__file__)
+sys.path.append(__dir__)
+sys.path.append(os.path.join(__dir__, '..'))
+
+import numpy as np
+from .locality_aware_nms import nms_locality
+# import lanms
+import cv2
+import time
+
+
+class SASTPostProcess(object):
+    """
+    The post process for SAST.
+    """
+
+    def __init__(self,
+                 score_thresh=0.5,
+                 nms_thresh=0.2,
+                 sample_pts_num=2,
+                 shrink_ratio_of_width=0.3,
+                 expand_scale=1.0,
+                 tcl_map_thresh=0.5,
+                 **kwargs):
+
+        self.score_thresh = score_thresh
+        self.nms_thresh = nms_thresh
+        self.sample_pts_num = sample_pts_num
+        self.shrink_ratio_of_width = shrink_ratio_of_width
+        self.expand_scale = expand_scale
+        self.tcl_map_thresh = tcl_map_thresh
+        
+        # c++ la-nms is faster, but only support python 3.5
+        self.is_python35 = False
+        if sys.version_info.major == 3 and sys.version_info.minor == 5:
+            self.is_python35 = True
+            
+    def point_pair2poly(self, point_pair_list):
+        """
+        Transfer vertical point_pairs into poly point in clockwise.
+        """
+        # constract poly
+        point_num = len(point_pair_list) * 2
+        point_list = [0] * point_num
+        for idx, point_pair in enumerate(point_pair_list):
+            point_list[idx] = point_pair[0]
+            point_list[point_num - 1 - idx] = point_pair[1]
+        return np.array(point_list).reshape(-1, 2)
+    
+    def shrink_quad_along_width(self, quad, begin_width_ratio=0., end_width_ratio=1.):
+        """ 
+        Generate shrink_quad_along_width.
+        """
+        ratio_pair = np.array([[begin_width_ratio], [end_width_ratio]], dtype=np.float32)
+        p0_1 = quad[0] + (quad[1] - quad[0]) * ratio_pair
+        p3_2 = quad[3] + (quad[2] - quad[3]) * ratio_pair
+        return np.array([p0_1[0], p0_1[1], p3_2[1], p3_2[0]])
+    
+    def expand_poly_along_width(self, poly, shrink_ratio_of_width=0.3):
+        """
+        expand poly along width.
+        """
+        point_num = poly.shape[0]
+        left_quad = np.array([poly[0], poly[1], poly[-2], poly[-1]], dtype=np.float32)
+        left_ratio = -shrink_ratio_of_width * np.linalg.norm(left_quad[0] - left_quad[3]) / \
+                    (np.linalg.norm(left_quad[0] - left_quad[1]) + 1e-6)
+        left_quad_expand = self.shrink_quad_along_width(left_quad, left_ratio, 1.0)
+        right_quad = np.array([poly[point_num // 2 - 2], poly[point_num // 2 - 1],
+                            poly[point_num // 2], poly[point_num // 2 + 1]], dtype=np.float32)
+        right_ratio = 1.0 + \
+                    shrink_ratio_of_width * np.linalg.norm(right_quad[0] - right_quad[3]) / \
+                    (np.linalg.norm(right_quad[0] - right_quad[1]) + 1e-6)
+        right_quad_expand = self.shrink_quad_along_width(right_quad, 0.0, right_ratio)
+        poly[0] = left_quad_expand[0]
+        poly[-1] = left_quad_expand[-1]
+        poly[point_num // 2 - 1] = right_quad_expand[1]
+        poly[point_num // 2] = right_quad_expand[2]
+        return poly
+
+    def restore_quad(self, tcl_map, tcl_map_thresh, tvo_map):
+        """Restore quad."""
+        xy_text = np.argwhere(tcl_map[:, :, 0] > tcl_map_thresh)
+        xy_text = xy_text[:, ::-1] # (n, 2)
+
+        # Sort the text boxes via the y axis
+        xy_text = xy_text[np.argsort(xy_text[:, 1])]
+
+        scores = tcl_map[xy_text[:, 1], xy_text[:, 0], 0]
+        scores = scores[:, np.newaxis]
+
+        # Restore
+        point_num = int(tvo_map.shape[-1] / 2)
+        assert point_num == 4
+        tvo_map = tvo_map[xy_text[:, 1], xy_text[:, 0], :]
+        xy_text_tile = np.tile(xy_text, (1, point_num)) # (n, point_num * 2)
+        quads = xy_text_tile - tvo_map
+
+        return scores, quads, xy_text
+
+    def quad_area(self, quad):
+        """
+        compute area of a quad.
+        """
+        edge = [
+            (quad[1][0] - quad[0][0]) * (quad[1][1] + quad[0][1]),
+            (quad[2][0] - quad[1][0]) * (quad[2][1] + quad[1][1]),
+            (quad[3][0] - quad[2][0]) * (quad[3][1] + quad[2][1]),
+            (quad[0][0] - quad[3][0]) * (quad[0][1] + quad[3][1])
+        ]
+        return np.sum(edge) / 2.
+        
+    def nms(self, dets):
+        if self.is_python35:
+            import lanms
+            dets = lanms.merge_quadrangle_n9(dets, self.nms_thresh)
+        else:
+            dets = nms_locality(dets, self.nms_thresh)
+        return dets
+
+    def cluster_by_quads_tco(self, tcl_map, tcl_map_thresh, quads, tco_map):
+        """
+        Cluster pixels in tcl_map based on quads.
+        """
+        instance_count = quads.shape[0] + 1 # contain background
+        instance_label_map = np.zeros(tcl_map.shape[:2], dtype=np.int32)
+        if instance_count == 1:
+            return instance_count, instance_label_map
+
+        # predict text center
+        xy_text = np.argwhere(tcl_map[:, :, 0] > tcl_map_thresh)
+        n = xy_text.shape[0]
+        xy_text = xy_text[:, ::-1] # (n, 2)
+        tco = tco_map[xy_text[:, 1], xy_text[:, 0], :] # (n, 2)
+        pred_tc = xy_text - tco
+        
+        # get gt text center
+        m = quads.shape[0]
+        gt_tc = np.mean(quads, axis=1) # (m, 2)
+
+        pred_tc_tile = np.tile(pred_tc[:, np.newaxis, :], (1, m, 1)) # (n, m, 2)
+        gt_tc_tile = np.tile(gt_tc[np.newaxis, :, :], (n, 1, 1)) # (n, m, 2)
+        dist_mat = np.linalg.norm(pred_tc_tile - gt_tc_tile, axis=2) # (n, m)
+        xy_text_assign = np.argmin(dist_mat, axis=1) + 1 # (n,)
+
+        instance_label_map[xy_text[:, 1], xy_text[:, 0]] = xy_text_assign
+        return instance_count, instance_label_map
+
+    def estimate_sample_pts_num(self, quad, xy_text):
+        """
+        Estimate sample points number.
+        """
+        eh = (np.linalg.norm(quad[0] - quad[3]) + np.linalg.norm(quad[1] - quad[2])) / 2.0
+        ew = (np.linalg.norm(quad[0] - quad[1]) + np.linalg.norm(quad[2] - quad[3])) / 2.0
+
+        dense_sample_pts_num = max(2, int(ew))
+        dense_xy_center_line = xy_text[np.linspace(0, xy_text.shape[0] - 1, dense_sample_pts_num,
+                                                endpoint=True, dtype=np.float32).astype(np.int32)]
+
+        dense_xy_center_line_diff = dense_xy_center_line[1:] - dense_xy_center_line[:-1]
+        estimate_arc_len = np.sum(np.linalg.norm(dense_xy_center_line_diff, axis=1))
+
+        sample_pts_num = max(2, int(estimate_arc_len / eh))
+        return sample_pts_num
+
+    def detect_sast(self, tcl_map, tvo_map, tbo_map, tco_map, ratio_w, ratio_h, src_w, src_h, 
+                shrink_ratio_of_width=0.3, tcl_map_thresh=0.5, offset_expand=1.0, out_strid=4.0):
+        """
+        first resize the tcl_map, tvo_map and tbo_map to the input_size, then restore the polys
+        """
+        # restore quad
+        scores, quads, xy_text = self.restore_quad(tcl_map, tcl_map_thresh, tvo_map)
+        dets = np.hstack((quads, scores)).astype(np.float32, copy=False)
+        dets = self.nms(dets)
+        if dets.shape[0] == 0:
+            return []
+        quads = dets[:, :-1].reshape(-1, 4, 2)
+
+        # Compute quad area
+        quad_areas = []
+        for quad in quads:
+            quad_areas.append(-self.quad_area(quad))
+
+        # instance segmentation
+        # instance_count, instance_label_map = cv2.connectedComponents(tcl_map.astype(np.uint8), connectivity=8)
+        instance_count, instance_label_map = self.cluster_by_quads_tco(tcl_map, tcl_map_thresh, quads, tco_map)
+
+        # restore single poly with tcl instance.
+        poly_list = []
+        for instance_idx in range(1, instance_count):
+            xy_text = np.argwhere(instance_label_map == instance_idx)[:, ::-1]
+            quad = quads[instance_idx - 1]
+            q_area = quad_areas[instance_idx - 1]
+            if q_area < 5:
+                continue
+            
+            #
+            len1 = float(np.linalg.norm(quad[0] -quad[1]))
+            len2 = float(np.linalg.norm(quad[1] -quad[2]))
+            min_len = min(len1, len2)
+            if min_len < 3:
+                continue
+
+            # filter small CC
+            if xy_text.shape[0] <= 0:
+                continue
+
+            # filter low confidence instance
+            xy_text_scores = tcl_map[xy_text[:, 1], xy_text[:, 0], 0] 
+            if np.sum(xy_text_scores) / quad_areas[instance_idx - 1] < 0.1:
+            # if np.sum(xy_text_scores) / quad_areas[instance_idx - 1] < 0.05:
+                continue
+
+            # sort xy_text
+            left_center_pt = np.array([[(quad[0, 0] + quad[-1, 0]) / 2.0,
+                                        (quad[0, 1] + quad[-1, 1]) / 2.0]]) # (1, 2)
+            right_center_pt = np.array([[(quad[1, 0] + quad[2, 0]) / 2.0,
+                                        (quad[1, 1] + quad[2, 1]) / 2.0]]) # (1, 2)
+            proj_unit_vec = (right_center_pt - left_center_pt) / \
+                            (np.linalg.norm(right_center_pt - left_center_pt) + 1e-6)
+            proj_value = np.sum(xy_text * proj_unit_vec, axis=1)
+            xy_text = xy_text[np.argsort(proj_value)]
+
+            # Sample pts in tcl map
+            if self.sample_pts_num == 0:
+                sample_pts_num = self.estimate_sample_pts_num(quad, xy_text)
+            else:
+                sample_pts_num = self.sample_pts_num
+            xy_center_line = xy_text[np.linspace(0, xy_text.shape[0] - 1, sample_pts_num,
+                                                endpoint=True, dtype=np.float32).astype(np.int32)]
+
+            point_pair_list = []
+            for x, y in xy_center_line:
+                # get corresponding offset
+                offset = tbo_map[y, x, :].reshape(2, 2)
+                if offset_expand != 1.0:
+                    offset_length = np.linalg.norm(offset, axis=1, keepdims=True)
+                    expand_length = np.clip(offset_length * (offset_expand - 1), a_min=0.5, a_max=3.0)
+                    offset_detal = offset / offset_length * expand_length
+                    offset = offset + offset_detal                
+                # original point
+                ori_yx = np.array([y, x], dtype=np.float32)
+                point_pair = (ori_yx +  offset)[:, ::-1]* out_strid / np.array([ratio_w, ratio_h]).reshape(-1, 2) 
+                point_pair_list.append(point_pair)
+
+            # ndarry: (x, 2), expand poly along width
+            detected_poly = self.point_pair2poly(point_pair_list)
+            detected_poly = self.expand_poly_along_width(detected_poly, shrink_ratio_of_width)
+            detected_poly[:, 0] = np.clip(detected_poly[:, 0], a_min=0, a_max=src_w)
+            detected_poly[:, 1] = np.clip(detected_poly[:, 1], a_min=0, a_max=src_h)
+            poly_list.append(detected_poly)
+
+        return poly_list
+
+    def __call__(self, outs_dict, shape_list):                
+        score_list = outs_dict['f_score']
+        border_list = outs_dict['f_border']
+        tvo_list = outs_dict['f_tvo']
+        tco_list = outs_dict['f_tco']
+                    
+        img_num = len(shape_list)
+        poly_lists = []
+        for ino in range(img_num):
+            p_score = score_list[ino].transpose((1,2,0)).numpy()
+            p_border = border_list[ino].transpose((1,2,0)).numpy()
+            p_tvo = tvo_list[ino].transpose((1,2,0)).numpy()
+            p_tco = tco_list[ino].transpose((1,2,0)).numpy()
+            src_h, src_w, ratio_h, ratio_w = shape_list[ino]
+
+            poly_list = self.detect_sast(p_score, p_tvo, p_border, p_tco, ratio_w, ratio_h, src_w, src_h, 
+                                         shrink_ratio_of_width=self.shrink_ratio_of_width, 
+                                         tcl_map_thresh=self.tcl_map_thresh, offset_expand=self.expand_scale)
+            poly_lists.append({'points': np.array(poly_list)})
+
+        return poly_lists
+

ppocr/utils/character.py

-# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-import numpy as np
-import string
-import re
-from .check import check_config_params
-import sys
-
-
-class CharacterOps(object):
-    """ Convert between text-label and text-index """
-
-    def __init__(self, config):
-        self.character_type = config['character_type']
-        self.loss_type = config['loss_type']
-        self.max_text_len = config['max_text_length']
-        if self.character_type == "en":
-            self.character_str = "0123456789abcdefghijklmnopqrstuvwxyz"
-            dict_character = list(self.character_str)
-        elif self.character_type == "ch":
-            character_dict_path = config['character_dict_path']
-            add_space = False
-            if 'use_space_char' in config:
-                add_space = config['use_space_char']
-            self.character_str = ""
-            with open(character_dict_path, "rb") as fin:
-                lines = fin.readlines()
-                for line in lines:
-                    line = line.decode('utf-8').strip("\n").strip("\r\n")
-                    self.character_str += line
-            if add_space:
-                self.character_str += " "
-            dict_character = list(self.character_str)
-        elif self.character_type == "en_sensitive":
-            # same with ASTER setting (use 94 char).
-            self.character_str = string.printable[:-6]
-            dict_character = list(self.character_str)
-        else:
-            self.character_str = None
-        assert self.character_str is not None, \
-            "Nonsupport type of the character: {}".format(self.character_str)
-        self.beg_str = "sos"
-        self.end_str = "eos"
-        if self.loss_type == "attention":
-            dict_character = [self.beg_str, self.end_str] + dict_character
-        elif self.loss_type == "srn":
-            dict_character = dict_character + [self.beg_str, self.end_str]
-        self.dict = {}
-        for i, char in enumerate(dict_character):
-            self.dict[char] = i
-        self.character = dict_character
-
-    def encode(self, text):
-        """convert text-label into text-index.
-        input:
-            text: text labels of each image. [batch_size]
-
-        output:
-            text: concatenated text index for CTCLoss.
-                    [sum(text_lengths)] = [text_index_0 + text_index_1 + ... + text_index_(n - 1)]
-            length: length of each text. [batch_size]
-        """
-        if self.character_type == "en":
-            text = text.lower()
-
-        text_list = []
-        for char in text:
-            if char not in self.dict:
-                continue
-            text_list.append(self.dict[char])
-        text = np.array(text_list)
-        return text
-
-    def decode(self, text_index, is_remove_duplicate=False):
-        """ convert text-index into text-label. """
-        char_list = []
-        char_num = self.get_char_num()
-
-        if self.loss_type == "attention":
-            beg_idx = self.get_beg_end_flag_idx("beg")
-            end_idx = self.get_beg_end_flag_idx("end")
-            ignored_tokens = [beg_idx, end_idx]
-        else:
-            ignored_tokens = [char_num]
-
-        for idx in range(len(text_index)):
-            if text_index[idx] in ignored_tokens:
-                continue
-            if is_remove_duplicate:
-                if idx > 0 and text_index[idx - 1] == text_index[idx]:
-                    continue
-            char_list.append(self.character[int(text_index[idx])])
-        text = ''.join(char_list)
-        return text
-
-    def get_char_num(self):
-        return len(self.character)
-
-    def get_beg_end_flag_idx(self, beg_or_end):
-        if self.loss_type == "attention":
-            if beg_or_end == "beg":
-                idx = np.array(self.dict[self.beg_str])
-            elif beg_or_end == "end":
-                idx = np.array(self.dict[self.end_str])
-            else:
-                assert False, "Unsupport type %s in get_beg_end_flag_idx"\
-                    % beg_or_end
-            return idx
-        else:
-            err = "error in get_beg_end_flag_idx when using the loss %s"\
-                % (self.loss_type)
-            assert False, err
-
-
-def cal_predicts_accuracy(char_ops,
-                          preds,
-                          preds_lod,
-                          labels,
-                          labels_lod,
-                          is_remove_duplicate=False):
-    acc_num = 0
-    img_num = 0
-    for ino in range(len(labels_lod) - 1):
-        beg_no = preds_lod[ino]
-        end_no = preds_lod[ino + 1]
-        preds_text = preds[beg_no:end_no].reshape(-1)
-        preds_text = char_ops.decode(preds_text, is_remove_duplicate)
-
-        beg_no = labels_lod[ino]
-        end_no = labels_lod[ino + 1]
-        labels_text = labels[beg_no:end_no].reshape(-1)
-        labels_text = char_ops.decode(labels_text, is_remove_duplicate)
-        img_num += 1
-
-        if preds_text == labels_text:
-            acc_num += 1
-    acc = acc_num * 1.0 / img_num
-    return acc, acc_num, img_num
-
-
-def cal_predicts_accuracy_srn(char_ops,
-                              preds,
-                              labels,
-                              max_text_len,
-                              is_debug=False):
-    acc_num = 0
-    img_num = 0
-
-    char_num = char_ops.get_char_num()
-
-    total_len = preds.shape[0]
-    img_num = int(total_len / max_text_len)
-    for i in range(img_num):
-        cur_label = []
-        cur_pred = []
-        for j in range(max_text_len):
-            if labels[j + i * max_text_len] != int(char_num-1):  #0
-                cur_label.append(labels[j + i * max_text_len][0])
-            else:
-                break
-
-        for j in range(max_text_len + 1):
-            if j < len(cur_label) and preds[j + i * max_text_len][
-                    0] != cur_label[j]:
-                break
-            elif j == len(cur_label) and j == max_text_len:
-                acc_num += 1
-                break
-            elif j == len(cur_label) and preds[j + i * max_text_len][0] == int(char_num-1):
-                acc_num += 1
-                break
-    acc = acc_num * 1.0 / img_num
-    return acc, acc_num, img_num
-
-
-def convert_rec_attention_infer_res(preds):
-    img_num = preds.shape[0]
-    target_lod = [0]
-    convert_ids = []
-    for ino in range(img_num):
-        end_pos = np.where(preds[ino, :] == 1)[0]
-        if len(end_pos) <= 1:
-            text_list = preds[ino, 1:]
-        else:
-            text_list = preds[ino, 1:end_pos[1]]
-        target_lod.append(target_lod[ino] + len(text_list))
-        convert_ids = convert_ids + list(text_list)
-    convert_ids = np.array(convert_ids)
-    convert_ids = convert_ids.reshape((-1, 1))
-    return convert_ids, target_lod
-
-
-def convert_rec_label_to_lod(ori_labels):
-    img_num = len(ori_labels)
-    target_lod = [0]
-    convert_ids = []
-    for ino in range(img_num):
-        target_lod.append(target_lod[ino] + len(ori_labels[ino]))
-        convert_ids = convert_ids + list(ori_labels[ino])
-    convert_ids = np.array(convert_ids)
-    convert_ids = convert_ids.reshape((-1, 1))
-    return convert_ids, target_lod

ppocr/utils/check.py

-# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-from __future__ import absolute_import
-from __future__ import division
-from __future__ import print_function
-from __future__ import unicode_literals
-
-import sys
-
-import logging
-logger = logging.getLogger(__name__)
-
-
-def check_config_params(config, config_name, params):
-    for param in params:
-        if param not in config:
-            err = "param %s didn't find in %s!" % (param, config_name)
-            assert False, err
-    return

ppocr/utils/dict/en_dict.txt

+0
+1
+2
+3
+4
+5
+6
+7
+8
+9
+a
+b
+c
+d
+e
+f
+g
+h
+i
+j
+k
+l
+m
+n
+o
+p
+q
+r
+s
+t
+u
+v
+w
+x
+y
+z
+A
+B
+C
+D
+E
+F
+G
+H
+I
+J
+K
+L
+M
+N
+O
+P
+Q
+R
+S
+T
+U
+V
+W
+X
+Y
+Z
+ 

ppocr/utils/dict/french_dict.txt

+f
+e
+n
+c
+h
+_
+i
+m
+g
+/
+r
+v
+a
+l
+t
+w
+o
+d
+6
+1
+.
+p
+B
+u
+2
+à
+3
+R
+y
+4
+U
+E
+A
+5
+P
+O
+S
+T
+D
+7
+Z
+8
+I
+N
+L
+G
+M
+H
+0
+J
+K
+-
+9
+F
+C
+V
+é
+X
+'
+s
+Q
+:
+è
+x
+b
+Y
+Œ
+É
+z
+W
+Ç
+È
+k
+Ô
+ô
+€
+À
+Ê
+q
+ù
+°
+ê
+î
+*
+Â
+j
+"
+,
+â
+%
+û
+ç
+ü
+?
+!
+;
+ö
+(
+)
+ï
+º
+ó
+ø
+å
++
+™
+á
+Ë
+<
+²
+Á
+Î
+&
+@
+œ
+ε
+Ü
+ë
+[
+]
+í
+ò
+Ö
+ä
+ß
+«
+»
+ú
+ñ
+æ
+µ
+³
+Å
+$
+#
+ 

ppocr/utils/dict/german_dict.txt

+!
+"
+$
+%
+&
+'
+(
+)
+*
++
+,
+-
+.
+/
+0
+1
+2
+3
+4
+5
+6
+7
+8
+9
+:
+;
+=
+>
+?
+@
+A
+B
+C
+D
+E
+F
+G
+H
+I
+J
+K
+L
+M
+N
+O
+P
+Q
+R
+S
+T
+U
+V
+W
+X
+Y
+Z
+[
+]
+_
+a
+b
+c
+d
+e
+f
+g
+h
+i
+j
+k
+l
+m
+n
+o
+p
+q
+r
+s
+t
+u
+v
+w
+x
+y
+z
+¡
+¢
+£
+¤
+¥
+¦
+§
+¨
+©
+ª
+«
+¬
+­
+®
+¯
+°
+±
+²
+³
+´
+µ
+¶
+·
+¸
+¹
+º
+»
+¼
+½
+¿
+Â
+Ã
+Å
+Ê
+Î
+Ð
+á
+â
+å
+æ
+é
+ 

ppocr/utils/dict/japan_dict.txt

+!
+"
+#
+$
+%
+&
+'
+(
+)
+*
++
+,
+-
+.
+/
+0
+1
+2
+3
+4
+5
+6
+7
+8
+9
+:
+;
+<
+=
+>
+?
+A
+B
+C
+D
+E
+F
+G
+H
+I
+J
+K
+L
+M
+N
+O
+P
+Q
+R
+S
+T
+U
+V
+W
+X
+Y
+Z
+[
+]
+_
+`
+a
+b
+c
+d
+e
+f
+g
+h
+i
+j
+k
+l
+m
+n
+o
+p
+q
+r
+s
+t
+u
+v
+w
+x
+y
+z
+©
+°
+²
+´
+½
+Á
+Ä
+Å
+Ç
+È
+É
+Í
+Ó
+Ö
+×
+Ü
+ß
+à
+á
+â
+ã
+ä
+å
+æ
+ç
+è
+é
+ê
+ë
+í
+ð
+ñ
+ò
+ó
+ô
+õ
+ö
+ø
+ú
+û
+ü
+ý
+ā
+ă
+ą
+ć
+Č
+č
+đ
+ē
+ė
+ę
+ğ
+ī
+ı
+Ł
+ł
+ń
+ň
+ō
+ř
+Ş
+ş
+Š
+š
+ţ
+ū
+ż
+Ž
+ž
+Ș
+ș
+ț
+Δ
+α
+λ
+μ
+φ
+Г
+О
+а
+в
+л
+о
+р
+с
+т
+я
+ồ
+​
+—
+―
+’
+“
+”
+…
+℃
+→
+∇
+−
+■
+☆
+　
+、
+。
+々
+〆
+〈
+〉
+「
+」
+『
+』
+〔
+〕
+〜
+ぁ
+あ
+ぃ
+い
+う
+ぇ
+え
+ぉ
+お
+か
+が
+き
+ぎ
+く
+ぐ
+け
+げ
+こ
+ご
+さ
+ざ
+し
+じ
+す
+ず
+せ
+ぜ
+そ
+ぞ
+た
+だ
+ち
+ぢ
+っ
+つ
+づ
+て
+で
+と
+ど
+な
+に
+ぬ
+ね
+の
+は
+ば
+ぱ
+ひ
+び
+ぴ
+ふ
+ぶ
+ぷ
+へ
+べ
+ぺ
+ほ
+ぼ
+ぽ
+ま
+み
+む
+め
+も
+ゃ
+や
+ゅ
+ゆ
+ょ
+よ
+ら
+り
+る
+れ
+ろ
+わ
+ゑ
+を
+ん
+ゝ
+ゞ
+ァ
+ア
+ィ
+イ
+ゥ
+ウ
+ェ
+エ
+ォ
+オ
+カ
+ガ
+キ
+ギ
+ク
+グ
+ケ
+ゲ
+コ
+ゴ
+サ
+ザ
+シ
+ジ
+ス
+ズ
+セ
+ゼ
+ソ
+ゾ
+タ
+ダ
+チ
+ヂ
+ッ
+ツ
+ヅ
+テ
+デ
+ト
+ド
+ナ
+ニ
+ヌ
+ネ
+ノ
+ハ
+バ
+パ
+ヒ
+ビ
+ピ
+フ
+ブ
+プ
+ヘ
+ベ
+ペ
+ホ
+ボ
+ポ
+マ
+ミ
+ム
+メ
+モ
+ャ
+ヤ
+ュ
+ユ
+ョ
+ヨ
+ラ
+リ
+ル
+レ
+ロ
+ワ
+ヰ
+ン
+ヴ
+ヵ
+ヶ
+・
+ー
+㈱
+一
+丁
+七
+万
+丈
+三
+上
+下
+不
+与
+丑
+且
+世
+丘
+丙
+丞
+両
+並
+中
+串
+丸
+丹
+主
+丼
+丿
+乃
+久
+之
+乎
+乏
+乗
+乘
+乙
+九
+乞
+也
+乱
+乳
+乾
+亀
+了
+予
+争
+事
+二
+于
+互
+五
+井
+亘
+亙
+些
+亜
+亟
+亡
+交
+亥
+亦
+亨
+享
+京
+亭
+亮
+人
+什
+仁
+仇
+今
+介
+仍
+仏
+仔
+仕
+他
+仗
+付
+仙
+代
+令
+以
+仮
+仰
+仲
+件
+任
+企
+伊
+伍
+伎
+伏
+伐
+休
+会
+伝
+伯
+估
+伴
+伶
+伸
+伺
+似
+伽
+佃
+但
+位
+低
+住
+佐
+佑
+体
+何
+余
+佚
+佛
+作
+佩
+佳
+併
+佶
+使
+侈
+例
+侍
+侏
+侑
+侘
+供
+依
+侠
+価
+侮
+侯
+侵
+侶
+便
+係
+促
+俄
+俊
+俔
+俗
+俘
+保
+信
+俣
+俤
+修
+俯
+俳
+俵
+俸
+俺
+倉
+個
+倍
+倒
+候
+借
+倣
+値
+倫
+倭
+倶
+倹
+偃
+假
+偈
+偉
+偏
+偐
+偕
+停
+健
+側
+偵
+偶
+偽
+傀
+傅
+傍
+傑
+傘
+備
+催
+傭
+傲
+傳
+債
+傷
+傾
+僊
+働
+像
+僑
+僕
+僚
+僧
+僭
+僮
+儀
+億
+儇
+儒
+儛
+償
+儡
+優
+儲
+儺
+儼
+兀
+允
+元
+兄
+充
+兆
+先
+光
+克
+兌
+免
+兎
+児
+党
+兜
+入
+全
+八
+公
+六
+共
+兵
+其
+具
+典
+兼
+内
+円
+冊
+再
+冑
+冒
+冗
+写
+冠
+冤
+冥
+冨
+冬
+冲
+决
+冶
+冷
+准
+凉
+凋
+凌
+凍
+凛
+凝
+凞
+几
+凡
+処
+凪
+凰
+凱
+凶
+凸
+凹
+出
+函
+刀
+刃
+分
+切
+刈
+刊
+刎
+刑
+列
+初
+判
+別
+利
+刪
+到
+制
+刷
+券
+刹
+刺
+刻
+剃
+則
+削
+剋
+前
+剖
+剛
+剣
+剤
+剥
+剪
+副
+剰
+割
+創
+剽
+劇
+劉
+劔
+力
+功
+加
+劣
+助
+努
+劫
+劭
+励
+労
+効
+劾
+勃
+勅
+勇
+勉
+勒
+動
+勘
+務
+勝
+募
+勢
+勤
+勧
+勲
+勺
+勾
+勿
+匁
+匂
+包
+匏
+化
+北
+匙
+匝
+匠
+匡
+匣
+匯
+匲
+匹
+区
+医
+匿
+十
+千
+升
+午
+卉
+半
+卍
+卑
+卒
+卓
+協
+南
+単
+博
+卜
+占
+卦
+卯
+印
+危
+即
+却
+卵
+卸
+卿
+厄
+厚
+原
+厠
+厨
+厩
+厭
+厳
+去
+参
+又
+叉
+及
+友
+双
+反
+収
+叔
+取
+受
+叙
+叛
+叟
+叡
+叢
+口
+古
+句
+叩
+只
+叫
+召
+可
+台
+叱
+史
+右
+叶
+号
+司
+吃
+各
+合
+吉
+吊
+同
+名
+后
+吏
+吐
+向
+君
+吝
+吟
+吠
+否
+含
+吸
+吹
+吻
+吽
+吾
+呂
+呆
+呈
+呉
+告
+呑
+周
+呪
+呰
+味
+呼
+命
+咀
+咄
+咋
+和
+咒
+咫
+咲
+咳
+咸
+哀
+品
+哇
+哉
+員
+哨
+哩
+哭
+哲
+哺
+唄
+唆
+唇
+唐
+唖
+唯
+唱
+唳
+唸
+唾
+啄
+商
+問
+啓
+啼
+善
+喋
+喚
+喜
+喝
+喧
+喩
+喪
+喫
+喬
+單
+喰
+営
+嗅
+嗇
+嗔
+嗚
+嗜
+嗣
+嘆
+嘉
+嘗
+嘘
+嘩
+嘯
+嘱
+嘲
+嘴
+噂
+噌
+噛
+器
+噴
+噺
+嚆
+嚢
+囀
+囃
+囉
+囚
+四
+回
+因
+団
+困
+囲
+図
+固
+国
+圀
+圃
+國
+圏
+園
+圓
+團
+圜
+土
+圧
+在
+圭
+地
+址
+坂
+均
+坊
+坐
+坑
+坡
+坤
+坦
+坪
+垂
+型
+垢
+垣
+埃
+埋
+城
+埒
+埔
+域
+埠
+埴
+埵
+執
+培
+基
+埼
+堀
+堂
+堅
+堆
+堕
+堤
+堪
+堯
+堰
+報
+場
+堵
+堺
+塀
+塁
+塊
+塑
+塔
+塗
+塘
+塙
+塚
+塞
+塩
+填
+塵
+塾
+境
+墉
+墓
+増
+墜
+墟
+墨
+墳
+墺
+墻
+墾
+壁
+壇
+壊
+壌
+壕
+士
+壬
+壮
+声
+壱
+売
+壷
+壹
+壺
+壽
+変
+夏
+夕
+外
+夙
+多
+夜
+夢
+夥
+大
+天
+太
+夫
+夬
+夭
+央
+失
+夷
+夾
+奄
+奇
+奈
+奉
+奎
+奏
+契
+奔
+奕
+套
+奘
+奠
+奢
+奥
+奨
+奪
+奮
+女
+奴
+奸
+好
+如
+妃
+妄
+妊
+妍
+妓
+妖
+妙
+妥
+妨
+妬
+妲
+妹
+妻
+妾
+姉
+始
+姐
+姓
+委
+姚
+姜
+姞
+姥
+姦
+姨
+姪
+姫
+姶
+姻
+姿
+威
+娑
+娘
+娟
+娠
+娩
+娯
+娼
+婆
+婉
+婚
+婢
+婦
+婬
+婿
+媄
+媒
+媓
+媚
+媛
+媞
+媽
+嫁
+嫄
+嫉
+嫌
+嫐
+嫗
+嫡
+嬉
+嬌
+嬢
+嬪
+嬬
+嬾
+孁
+子
+孔
+字
+存
+孚
+孝
+孟
+季
+孤
+学
+孫
+孵
+學
+宅
+宇
+守
+安
+宋
+完
+宍
+宏
+宕
+宗
+官
+宙
+定
+宛
+宜
+宝
+実
+客
+宣
+室
+宥
+宮
+宰
+害
+宴
+宵
+家
+宸
+容
+宿
+寂
+寄
+寅
+密
+寇
+富
+寒
+寓
+寔
+寛
+寝
+察
+寡
+實
+寧
+審
+寮
+寵
+寶
+寸
+寺
+対
+寿
+封
+専
+射
+将
+尉
+尊
+尋
+對
+導
+小
+少
+尖
+尚
+尤
+尪
+尭
+就
+尹
+尺
+尻
+尼
+尽
+尾
+尿
+局
+居
+屈
+届
+屋
+屍
+屎
+屏
+屑
+屓
+展
+属
+屠
+層
+履
+屯
+山
+岐
+岑
+岡
+岩
+岫
+岬
+岳
+岷
+岸
+峠
+峡
+峨
+峯
+峰
+島
+峻
+崇
+崋
+崎
+崑
+崖
+崗
+崛
+崩
+嵌
+嵐
+嵩
+嵯
+嶂
+嶋
+嶠
+嶺
+嶼
+嶽
+巀
+巌
+巒
+巖
+川
+州
+巡
+巣
+工
+左
+巧
+巨
+巫
+差
+己
+巳
+巴
+巷
+巻
+巽
+巾
+市
+布
+帆
+希
+帖
+帚
+帛
+帝
+帥
+師
+席
+帯
+帰
+帳
+帷
+常
+帽
+幄
+幅
+幇
+幌
+幔
+幕
+幟
+幡
+幢
+幣
+干
+平
+年
+并
+幸
+幹
+幻
+幼
+幽
+幾
+庁
+広
+庄
+庇
+床
+序
+底
+庖
+店
+庚
+府
+度
+座
+庫
+庭
+庵
+庶
+康
+庸
+廂
+廃
+廉
+廊
+廓
+廟
+廠
+廣
+廬
+延
+廷
+建
+廻
+廼
+廿
+弁
+弄
+弉
+弊
+弌
+式
+弐
+弓
+弔
+引
+弖
+弗
+弘
+弛
+弟
+弥
+弦
+弧
+弱
+張
+強
+弼
+弾
+彈
+彊
+彌
+彎
+当
+彗
+彙
+彝
+形
+彦
+彩
+彫
+彬
+彭
+彰
+影
+彷
+役
+彼
+往
+征
+徂
+径
+待
+律
+後
+徐
+徑
+徒
+従
+得
+徠
+御
+徧
+徨
+復
+循
+徭
+微
+徳
+徴
+德
+徹
+徽
+心
+必
+忉
+忌
+忍
+志
+忘
+忙
+応
+忠
+快
+忯
+念
+忻
+忽
+忿
+怒
+怖
+思
+怠
+怡
+急
+性
+怨
+怪
+怯
+恂
+恋
+恐
+恒
+恕
+恣
+恤
+恥
+恨
+恩
+恬
+恭
+息
+恵
+悉
+悌
+悍
+悔
+悟
+悠
+患
+悦
+悩
+悪
+悲
+悼
+情
+惇
+惑
+惚
+惜
+惟
+惠
+惣
+惧
+惨
+惰
+想
+惹
+惺
+愈
+愉
+愍
+意
+愔
+愚
+愛
+感
+愷
+愿
+慈
+態
+慌
+慎
+慕
+慢
+慣
+慧
+慨
+慮
+慰
+慶
+憂
+憎
+憐
+憑
+憙
+憤
+憧
+憩
+憬
+憲
+憶
+憾
+懇
+應
+懌
+懐
+懲
+懸
+懺
+懽
+懿
+戈
+戊
+戌
+戎
+成
+我
+戒
+戔
+或
+戚
+戟
+戦
+截
+戮
+戯
+戴
+戸
+戻
+房
+所
+扁
+扇
+扈
+扉
+手
+才
+打
+払
+托
+扮
+扱
+扶
+批
+承
+技
+抄
+把
+抑
+抓
+投
+抗
+折
+抜
+択
+披
+抱
+抵
+抹
+押
+抽
+担
+拇
+拈
+拉
+拍
+拏
+拐
+拒
+拓
+拘
+拙
+招
+拝
+拠
+拡
+括
+拭
+拳
+拵
+拶
+拾
+拿
+持
+挂
+指
+按
+挑
+挙
+挟
+挨
+振
+挺
+挽
+挿
+捉
+捕
+捗
+捜
+捧
+捨
+据
+捺
+捻
+掃
+掄
+授
+掌
+排
+掖
+掘
+掛
+掟
+採
+探
+掣
+接
+控
+推
+掩
+措
+掬
+掲
+掴
+掻
+掾
+揃
+揄
+揆
+揉
+描
+提
+揖
+揚
+換
+握
+揮
+援
+揶
+揺
+損
+搦
+搬
+搭
+携
+搾
+摂
+摘
+摩
+摸
+摺
+撃
+撒
+撞
+撤
+撥
+撫
+播
+撮
+撰
+撲
+撹
+擁
+操
+擔
+擦
+擬
+擾
+攘
+攝
+攣
+支
+收
+改
+攻
+放
+政
+故
+敏
+救
+敗
+教
+敢
+散
+敦
+敬
+数
+整
+敵
+敷
+斂
+文
+斉
+斎
+斐
+斑
+斗
+料
+斜
+斟
+斤
+斥
+斧
+斬
+断
+斯
+新
+方
+於
+施
+旁
+旅
+旋
+旌
+族
+旗
+旛
+无
+旡
+既
+日
+旦
+旧
+旨
+早
+旬
+旭
+旺
+旻
+昂
+昆
+昇
+昉
+昌
+明
+昏
+易
+昔
+星
+映
+春
+昧
+昨
+昪
+昭
+是
+昵
+昼
+晁
+時
+晃
+晋
+晏
+晒
+晟
+晦
+晧
+晩
+普
+景
+晴
+晶
+智
+暁
+暇
+暈
+暉
+暑
+暖
+暗
+暘
+暢
+暦
+暫
+暮
+暲
+暴
+暹
+暾
+曄
+曇
+曉
+曖
+曙
+曜
+曝
+曠
+曰
+曲
+曳
+更
+書
+曹
+曼
+曽
+曾
+替
+最
+會
+月
+有
+朋
+服
+朏
+朔
+朕
+朗
+望
+朝
+期
+朧
+木
+未
+末
+本
+札
+朱
+朴
+机
+朽
+杁
+杉
+李
+杏
+材
+村
+杓
+杖
+杜
+杞
+束
+条
+杢
+杣
+来
+杭
+杮
+杯
+東
+杲
+杵
+杷
+杼
+松
+板
+枅
+枇
+析
+枓
+枕
+林
+枚
+果
+枝
+枠
+枡
+枢
+枯
+枳
+架
+柄
+柊
+柏
+某
+柑
+染
+柔
+柘
+柚
+柯
+柱
+柳
+柴
+柵
+査
+柾
+柿
+栂
+栃
+栄
+栖
+栗
+校
+株
+栲
+栴
+核
+根
+栻
+格
+栽
+桁
+桂
+桃
+框
+案
+桐
+桑
+桓
+桔
+桜
+桝
+桟
+桧
+桴
+桶
+桾
+梁
+梅
+梆
+梓
+梔
+梗
+梛
+條
+梟
+梢
+梧
+梨
+械
+梱
+梲
+梵
+梶
+棄
+棋
+棒
+棗
+棘
+棚
+棟
+棠
+森
+棲
+棹
+棺
+椀
+椅
+椋
+植
+椎
+椏
+椒
+椙
+検
+椥
+椹
+椿
+楊
+楓
+楕
+楚
+楞
+楠
+楡
+楢
+楨
+楪
+楫
+業
+楮
+楯
+楳
+極
+楷
+楼
+楽
+概
+榊
+榎
+榕
+榛
+榜
+榮
+榱
+榴
+槃
+槇
+槊
+構
+槌
+槍
+槐
+様
+槙
+槻
+槽
+槿
+樂
+樋
+樓
+樗
+標
+樟
+模
+権
+横
+樫
+樵
+樹
+樺
+樽
+橇
+橋
+橘
+機
+橿
+檀
+檄
+檎
+檐
+檗
+檜
+檣
+檥
+檬
+檮
+檸
+檻
+櫃
+櫓
+櫛
+櫟
+櫨
+櫻
+欄
+欅
+欠
+次
+欣
+欧
+欲
+欺
+欽
+款
+歌
+歎
+歓
+止
+正
+此
+武
+歩
+歪
+歯
+歳
+歴
+死
+殆
+殉
+殊
+残
+殖
+殯
+殴
+段
+殷
+殺
+殻
+殿
+毀
+毅
+母
+毎
+毒
+比
+毘
+毛
+毫
+毬
+氈
+氏
+民
+気
+水
+氷
+永
+氾
+汀
+汁
+求
+汎
+汐
+汗
+汚
+汝
+江
+池
+汪
+汰
+汲
+決
+汽
+沂
+沃
+沅
+沆
+沈
+沌
+沐
+沓
+沖
+沙
+没
+沢
+沱
+河
+沸
+油
+治
+沼
+沽
+沿
+況
+泉
+泊
+泌
+法
+泗
+泡
+波
+泣
+泥
+注
+泯
+泰
+泳
+洋
+洒
+洗
+洛
+洞
+津
+洩
+洪
+洲
+洸
+洹
+活
+洽
+派
+流
+浄
+浅
+浙
+浚
+浜
+浣
+浦
+浩
+浪
+浮
+浴
+海
+浸
+涅
+消
+涌
+涙
+涛
+涯
+液
+涵
+涼
+淀
+淄
+淆
+淇
+淋
+淑
+淘
+淡
+淤
+淨
+淫
+深
+淳
+淵
+混
+淹
+添
+清
+済
+渉
+渋
+渓
+渕
+渚
+減
+渟
+渠
+渡
+渤
+渥
+渦
+温
+渫
+測
+港
+游
+渾
+湊
+湖
+湘
+湛
+湧
+湫
+湯
+湾
+湿
+満
+源
+準
+溜
+溝
+溢
+溥
+溪
+溶
+溺
+滄
+滅
+滋
+滌
+滑
+滕
+滝
+滞
+滴
+滸
+滹
+滿
+漁
+漂
+漆
+漉
+漏
+漑
+演
+漕
+漠
+漢
+漣
+漫
+漬
+漱
+漸
+漿
+潅
+潔
+潙
+潜
+潟
+潤
+潭
+潮
+潰
+潴
+澁
+澂
+澄
+澎
+澗
+澤
+澪
+澱
+澳
+激
+濁
+濃
+濟
+濠
+濡
+濤
+濫
+濯
+濱
+濾
+瀉
+瀋
+瀑
+瀕
+瀞
+瀟
+瀧
+瀬
+瀾
+灌
+灑
+灘
+火
+灯
+灰
+灸
+災
+炉
+炊
+炎
+炒
+炭
+炮
+炷
+点
+為
+烈
+烏
+烙
+烝
+烹
+焔
+焙
+焚
+無
+焦
+然
+焼
+煇
+煉
+煌
+煎
+煕
+煙
+煤
+煥
+照
+煩
+煬
+煮
+煽
+熈
+熊
+熙
+熟
+熨
+熱
+熹
+熾
+燃
+燈
+燎
+燔
+燕
+燗
+燥
+燭
+燻
+爆
+爐
+爪
+爬
+爲
+爵
+父
+爺
+爼
+爽
+爾
+片
+版
+牌
+牒
+牘
+牙
+牛
+牝
+牟
+牡
+牢
+牧
+物
+牲
+特
+牽
+犂
+犠
+犬
+犯
+状
+狂
+狄
+狐
+狗
+狙
+狛
+狡
+狩
+独
+狭
+狷
+狸
+狼
+猊
+猛
+猟
+猥
+猨
+猩
+猪
+猫
+献
+猴
+猶
+猷
+猾
+猿
+獄
+獅
+獏
+獣
+獲
+玄
+玅
+率
+玉
+王
+玖
+玩
+玲
+珀
+珂
+珈
+珉
+珊
+珍
+珎
+珞
+珠
+珣
+珥
+珪
+班
+現
+球
+理
+琉
+琢
+琥
+琦
+琮
+琲
+琳
+琴
+琵
+琶
+瑁
+瑋
+瑙
+瑚
+瑛
+瑜
+瑞
+瑠
+瑤
+瑩
+瑪
+瑳
+瑾
+璃
+璋
+璜
+璞
+璧
+璨
+環
+璵
+璽
+璿
+瓊
+瓔
+瓜
+瓢
+瓦
+瓶
+甍
+甑
+甕
+甘
+甚
+甞
+生
+産
+甥
+用
+甫
+田
+由
+甲
+申
+男
+町
+画
+界
+畏
+畑
+畔
+留
+畜
+畝
+畠
+畢
+略
+番
+異
+畳
+當
+畷
+畸
+畺
+畿
+疆
+疇
+疋
+疎
+疏
+疑
+疫
+疱
+疲
+疹
+疼
+疾
+病
+症
+痒
+痔
+痕
+痘
+痙
+痛
+痢
+痩
+痴
+痺
+瘍
+瘡
+瘧
+療
+癇
+癌
+癒
+癖
+癡
+癪
+発
+登
+白
+百
+的
+皆
+皇
+皋
+皐
+皓
+皮
+皺
+皿
+盂
+盃
+盆
+盈
+益
+盒
+盗
+盛
+盞
+盟
+盡
+監
+盤
+盥
+盧
+目
+盲
+直
+相
+盾
+省
+眉
+看
+県
+眞
+真
+眠
+眷
+眺
+眼
+着
+睡
+督
+睦
+睨
+睿
+瞋
+瞑
+瞞
+瞬
+瞭
+瞰
+瞳
+瞻
+瞼
+瞿
+矍
+矛
+矜
+矢
+知
+矧
+矩
+短
+矮
+矯
+石
+砂
+砌
+研
+砕
+砥
+砦
+砧
+砲
+破
+砺
+硝
+硫
+硬
+硯
+碁
+碇
+碌
+碑
+碓
+碕
+碗
+碣
+碧
+碩
+確
+碾
+磁
+磐
+磔
+磧
+磨
+磬
+磯
+礁
+礎
+礒
+礙
+礫
+礬
+示
+礼
+社
+祀
+祁
+祇
+祈
+祉
+祐
+祓
+祕
+祖
+祗
+祚
+祝
+神
+祟
+祠
+祢
+祥
+票
+祭
+祷
+祺
+禁
+禄
+禅
+禊
+禍
+禎
+福
+禔
+禖
+禛
+禦
+禧
+禮
+禰
+禹
+禽
+禿
+秀
+私
+秋
+科
+秒
+秘
+租
+秤
+秦
+秩
+称
+移
+稀
+程
+税
+稔
+稗
+稙
+稚
+稜
+稠
+種
+稱
+稲
+稷
+稻
+稼
+稽
+稿
+穀
+穂
+穆
+積
+穎
+穏
+穗
+穜
+穢
+穣
+穫
+穴
+究
+空
+突
+窃
+窄
+窒
+窓
+窟
+窠
+窩
+窪
+窮
+窯
+竃
+竄
+竈
+立
+站
+竜
+竝
+竟
+章
+童
+竪
+竭
+端
+竴
+競
+竹
+竺
+竽
+竿
+笄
+笈
+笏
+笑
+笙
+笛
+笞
+笠
+笥
+符
+第
+笹
+筅
+筆
+筇
+筈
+等
+筋
+筌
+筍
+筏
+筐
+筑
+筒
+答
+策
+筝
+筥
+筧
+筬
+筮
+筯
+筰
+筵
+箆
+箇
+箋
+箏
+箒
+箔
+箕
+算
+箙
+箜
+管
+箪
+箭
+箱
+箸
+節
+篁
+範
+篆
+篇
+築
+篋
+篌
+篝
+篠
+篤
+篥
+篦
+篩
+篭
+篳
+篷
+簀
+簒
+簡
+簧
+簪
+簫
+簺
+簾
+簿
+籀
+籃
+籌
+籍
+籐
+籟
+籠
+籤
+籬
+米
+籾
+粂
+粉
+粋
+粒
+粕
+粗
+粘
+粛
+粟
+粥
+粧
+粮
+粳
+精
+糊
+糖
+糜
+糞
+糟
+糠
+糧
+糯
+糸
+糺
+系
+糾
+紀
+約
+紅
+紋
+納
+紐
+純
+紗
+紘
+紙
+級
+紛
+素
+紡
+索
+紫
+紬
+累
+細
+紳
+紵
+紹
+紺
+絁
+終
+絃
+組
+絅
+経
+結
+絖
+絞
+絡
+絣
+給
+統
+絲
+絵
+絶
+絹
+絽
+綏
+經
+継
+続
+綜
+綟
+綬
+維
+綱
+網
+綴
+綸
+綺
+綽
+綾
+綿
+緊
+緋
+総
+緑
+緒
+線
+締
+緥
+編
+緩
+緬
+緯
+練
+緻
+縁
+縄
+縅
+縒
+縛
+縞
+縢
+縣
+縦
+縫
+縮
+縹
+總
+績
+繁
+繊
+繋
+繍
+織
+繕
+繝
+繦
+繧
+繰
+繹
+繼
+纂
+纈
+纏
+纐
+纒
+纛
+缶
+罔
+罠
+罧
+罪
+置
+罰
+署
+罵
+罷
+罹
+羂
+羅
+羆
+羇
+羈
+羊
+羌
+美
+群
+羨
+義
+羯
+羲
+羹
+羽
+翁
+翅
+翌
+習
+翔
+翛
+翠
+翡
+翫
+翰
+翺
+翻
+翼
+耀
+老
+考
+者
+耆
+而
+耐
+耕
+耗
+耨
+耳
+耶
+耽
+聊
+聖
+聘
+聚
+聞
+聟
+聡
+聨
+聯
+聰
+聲
+聴
+職
+聾
+肄
+肆
+肇
+肉
+肋
+肌
+肖
+肘
+肛
+肝
+股
+肢
+肥
+肩
+肪
+肯
+肱
+育
+肴
+肺
+胃
+胆
+背
+胎
+胖
+胚
+胝
+胞
+胡
+胤
+胱
+胴
+胸
+能
+脂
+脅
+脆
+脇
+脈
+脊
+脚
+脛
+脩
+脱
+脳
+腋
+腎
+腐
+腑
+腔
+腕
+腫
+腰
+腱
+腸
+腹
+腺
+腿
+膀
+膏
+膚
+膜
+膝
+膠
+膣
+膨
+膩
+膳
+膵
+膾
+膿
+臂
+臆
+臈
+臍
+臓
+臘
+臚
+臣
+臥
+臨
+自
+臭
+至
+致
+臺
+臼
+舂
+舅
+與
+興
+舌
+舍
+舎
+舒
+舖
+舗
+舘
+舜
+舞
+舟
+舩
+航
+般
+舳
+舶
+船
+艇
+艘
+艦
+艮
+良
+色
+艶
+芋
+芒
+芙
+芝
+芥
+芦
+芬
+芭
+芯
+花
+芳
+芸
+芹
+芻
+芽
+芿
+苅
+苑
+苔
+苗
+苛
+苞
+苡
+若
+苦
+苧
+苫
+英
+苴
+苻
+茂
+范
+茄
+茅
+茎
+茗
+茘
+茜
+茨
+茲
+茵
+茶
+茸
+茹
+草
+荊
+荏
+荒
+荘
+荷
+荻
+荼
+莞
+莪
+莫
+莬
+莱
+莵
+莽
+菅
+菊
+菌
+菓
+菖
+菘
+菜
+菟
+菩
+菫
+華
+菱
+菴
+萄
+萊
+萌
+萍
+萎
+萠
+萩
+萬
+萱
+落
+葉
+著
+葛
+葡
+董
+葦
+葩
+葬
+葭
+葱
+葵
+葺
+蒋
+蒐
+蒔
+蒙
+蒟
+蒡
+蒲
+蒸
+蒻
+蒼
+蒿
+蓄
+蓆
+蓉
+蓋
+蓑
+蓬
+蓮
+蓼
+蔀
+蔑
+蔓
+蔚
+蔡
+蔦
+蔬
+蔭
+蔵
+蔽
+蕃
+蕉
+蕊
+蕎
+蕨
+蕩
+蕪
+蕭
+蕾
+薄
+薇
+薊
+薔
+薗
+薙
+薛
+薦
+薨
+薩
+薪
+薫
+薬
+薭
+薮
+藁
+藉
+藍
+藏
+藐
+藝
+藤
+藩
+藪
+藷
+藹
+藺
+藻
+蘂
+蘆
+蘇
+蘊
+蘭
+虎
+虐
+虔
+虚
+虜
+虞
+號
+虫
+虹
+虻
+蚊
+蚕
+蛇
+蛉
+蛍
+蛎
+蛙
+蛛
+蛟
+蛤
+蛭
+蛮
+蛸
+蛹
+蛾
+蜀
+蜂
+蜃
+蜆
+蜊
+蜘
+蜜
+蜷
+蜻
+蝉
+蝋
+蝕
+蝙
+蝠
+蝦
+蝶
+蝿
+螂
+融
+螣
+螺
+蟄
+蟇
+蟠
+蟷
+蟹
+蟻
+蠢
+蠣
+血
+衆
+行
+衍
+衒
+術
+街
+衙
+衛
+衝
+衞
+衡
+衢
+衣
+表
+衫
+衰
+衵
+衷
+衽
+衾
+衿
+袁
+袈
+袋
+袍
+袒
+袖
+袙
+袞
+袢
+被
+袰
+袱
+袴
+袷
+袿
+裁
+裂
+裃
+装
+裏
+裔
+裕
+裘
+裙
+補
+裟
+裡
+裲
+裳
+裴
+裸
+裹
+製
+裾
+褂
+褄
+複
+褌
+褐
+褒
+褥
+褪
+褶
+褻
+襄
+襖
+襞
+襟
+襠
+襦
+襪
+襲
+襴
+襷
+西
+要
+覆
+覇
+覈
+見
+規
+視
+覗
+覚
+覧
+親
+覲
+観
+覺
+觀
+角
+解
+触
+言
+訂
+計
+討
+訓
+託
+記
+訛
+訟
+訢
+訥
+訪
+設
+許
+訳
+訴
+訶
+診
+註
+証
+詐
+詔
+評
+詛
+詞
+詠
+詢
+詣
+試
+詩
+詫
+詮
+詰
+話
+該
+詳
+誄
+誅
+誇
+誉
+誌
+認
+誓
+誕
+誘
+語
+誠
+誡
+誣
+誤
+誥
+誦
+説
+読
+誰
+課
+誼
+誾
+調
+談
+請
+諌
+諍
+諏
+諒
+論
+諚
+諜
+諟
+諡
+諦
+諧
+諫
+諭
+諮
+諱
+諶
+諷
+諸
+諺
+諾
+謀
+謄
+謌
+謎
+謗
+謙
+謚
+講
+謝
+謡
+謫
+謬
+謹
+證
+識
+譚
+譛
+譜
+警
+譬
+譯
+議
+譲
+譴
+護
+讀
+讃
+讐
+讒
+谷
+谿
+豅
+豆
+豊
+豎
+豐
+豚
+象
+豪
+豫
+豹
+貌
+貝
+貞
+負
+財
+貢
+貧
+貨
+販
+貪
+貫
+責
+貯
+貰
+貴
+買
+貸
+費
+貼
+貿
+賀
+賁
+賂
+賃
+賄
+資
+賈
+賊
+賎
+賑
+賓
+賛
+賜
+賞
+賠
+賢
+賣
+賤
+賦
+質
+賭
+購
+賽
+贄
+贅
+贈
+贋
+贔
+贖
+赤
+赦
+走
+赴
+起
+超
+越
+趙
+趣
+足
+趺
+趾
+跋
+跏
+距
+跡
+跨
+跪
+路
+跳
+践
+踊
+踏
+踐
+踞
+踪
+踵
+蹄
+蹉
+蹊
+蹟
+蹲
+蹴
+躅
+躇
+躊
+躍
+躑
+躙
+躪
+身
+躬
+躯
+躰
+車
+軋
+軌
+軍
+軒
+軟
+転
+軸
+軻
+軽
+軾
+較
+載
+輌
+輔
+輜
+輝
+輦
+輩
+輪
+輯
+輸
+輿
+轄
+轍
+轟
+轢
+辛
+辞
+辟
+辥
+辦
+辨
+辰
+辱
+農
+辺
+辻
+込
+迂
+迅
+迎
+近
+返
+迢
+迦
+迪
+迫
+迭
+述
+迷
+迹
+追
+退
+送
+逃
+逅
+逆
+逍
+透
+逐
+逓
+途
+逕
+逗
+這
+通
+逝
+逞
+速
+造
+逢
+連
+逮
+週
+進
+逸
+逼
+遁
+遂
+遅
+遇
+遊
+運
+遍
+過
+遐
+道
+達
+違
+遙
+遜
+遠
+遡
+遣
+遥
+適
+遭
+遮
+遯
+遵
+遷
+選
+遺
+遼
+避
+邀
+邁
+邂
+邃
+還
+邇
+邉
+邊
+邑
+那
+邦
+邨
+邪
+邯
+邵
+邸
+郁
+郊
+郎
+郡
+郢
+部
+郭
+郴
+郵
+郷
+都
+鄂
+鄙
+鄭
+鄰
+鄲
+酉
+酋
+酌
+配
+酎
+酒
+酔
+酢
+酥
+酪
+酬
+酵
+酷
+酸
+醍
+醐
+醒
+醗
+醜
+醤
+醪
+醵
+醸
+采
+釈
+釉
+釋
+里
+重
+野
+量
+釐
+金
+釘
+釜
+針
+釣
+釧
+釿
+鈍
+鈎
+鈐
+鈔
+鈞
+鈦
+鈴
+鈷
+鈸
+鈿
+鉄
+鉇
+鉉
+鉋
+鉛
+鉢
+鉤
+鉦
+鉱
+鉾
+銀
+銃
+銅
+銈
+銑
+銕
+銘
+銚
+銜
+銭
+鋏
+鋒
+鋤
+鋭
+鋲
+鋳
+鋸
+鋺
+鋼
+錆
+錍
+錐
+錘
+錠
+錣
+錦
+錫
+錬
+錯
+録
+錵
+鍋
+鍍
+鍑
+鍔
+鍛
+鍬
+鍮
+鍵
+鍼
+鍾
+鎌
+鎖
+鎗
+鎚
+鎧
+鎬
+鎮
+鎰
+鎹
+鏃
+鏑
+鏡
+鐃
+鐇
+鐐
+鐔
+鐘
+鐙
+鐚
+鐡
+鐵
+鐸
+鑁
+鑊
+鑑
+鑒
+鑚
+鑠
+鑢
+鑰
+鑵
+鑷
+鑼
+鑽
+鑿
+長
+門
+閃
+閇
+閉
+開
+閏
+閑
+間
+閔
+閘
+関
+閣
+閤
+閥
+閦
+閨
+閬
+閲
+閻
+閼
+閾
+闇
+闍
+闔
+闕
+闘
+關
+闡
+闢
+闥
+阜
+阪
+阮
+阯
+防
+阻
+阿
+陀
+陂
+附
+陌
+降
+限
+陛
+陞
+院
+陣
+除
+陥
+陪
+陬
+陰
+陳
+陵
+陶
+陸
+険
+陽
+隅
+隆
+隈
+隊
+隋
+階
+随
+隔
+際
+障
+隠
+隣
+隧
+隷
+隻
+隼
+雀
+雁
+雄
+雅
+集
+雇
+雉
+雊
+雋
+雌
+雍
+雑
+雖
+雙
+雛
+離
+難
+雨
+雪
+雫
+雰
+雲
+零
+雷
+雹
+電
+需
+震
+霊
+霍
+霖
+霜
+霞
+霧
+霰
+露
+靈
+青
+靖
+静
+靜
+非
+面
+革
+靫
+靭
+靱
+靴
+靺
+鞁
+鞄
+鞆
+鞋
+鞍
+鞏
+鞘
+鞠
+鞨
+鞭
+韋
+韓
+韜
+韮
+音
+韶
+韻
+響
+頁
+頂
+頃
+項
+順
+須
+頌
+預
+頑
+頒
+頓
+領
+頚
+頬
+頭
+頴
+頸
+頻
+頼
+顆
+題
+額
+顎
+顔
+顕
+顗
+願
+顛
+類
+顧
+顯
+風
+飛
+食
+飢
+飩
+飫
+飯
+飲
+飴
+飼
+飽
+飾
+餃
+餅
+餉
+養
+餌
+餐
+餓
+餘
+餝
+餡
+館
+饂
+饅
+饉
+饋
+饌
+饒
+饗
+首
+馗
+香
+馨
+馬
+馳
+馴
+駄
+駅
+駆
+駈
+駐
+駒
+駕
+駝
+駿
+騁
+騎
+騏
+騒
+験
+騙
+騨
+騰
+驕
+驚
+驛
+驢
+骨
+骸
+髄
+體
+高
+髙
+髢
+髪
+髭
+髮
+髷
+髻
+鬘
+鬚
+鬢
+鬨
+鬯
+鬱
+鬼
+魁
+魂
+魄
+魅
+魏
+魔
+魚
+魯
+鮎
+鮑
+鮒
+鮪
+鮫
+鮭
+鮮
+鯉
+鯔
+鯖
+鯛
+鯨
+鯰
+鯱
+鰐
+鰒
+鰭
+鰯
+鰰
+鰹
+鰻
+鱈
+鱒
+鱗
+鱧
+鳥
+鳩
+鳰
+鳳
+鳴
+鳶
+鴈
+鴉
+鴎
+鴛
+鴟
+鴦
+鴨
+鴫
+鴻
+鵄
+鵜
+鵞
+鵡
+鵬
+鵲
+鵺
+鶉
+鶏
+鶯
+鶴
+鷄
+鷙
+鷲
+鷹
+鷺
+鸚
+鸞
+鹸
+鹽
+鹿
+麁
+麒
+麓
+麗
+麝
+麞
+麟
+麦
+麩
+麹
+麺
+麻
+麾
+麿
+黄
+黌
+黍
+黒
+黙
+黛
+黠
+鼈
+鼉
+鼎
+鼓
+鼠
+鼻
+齊
+齋
+齟
+齢
+齬
+龍
+龕
+龗
+！
+＃
+％
+＆
+（
+）
+＋
+，
+－
+．
+／
+０
+１
+２
+３
+４
+５
+６
+７
+８
+９
+：
+；
+＝
+？
+＠
+Ａ
+Ｂ
+Ｃ
+Ｄ
+Ｅ
+Ｆ
+Ｇ
+Ｈ
+Ｉ
+Ｊ
+Ｋ
+Ｌ
+Ｍ
+Ｎ
+Ｏ
+Ｐ
+Ｒ
+Ｓ
+Ｔ
+Ｕ
+Ｖ
+Ｗ
+Ｘ
+Ｚ
+ａ
+ｃ
+ｄ
+ｅ
+ｆ
+ｈ
+ｉ
+ｊ
+ｋ
+ｌ
+ｍ
+ｎ
+ｏ
+ｐ
+ｒ
+ｓ
+ｔ
+ｕ
+ｙ
+ｚ
+～
+･
+ 

ppocr/utils/dict/korean_dict.txt

+!
+"
+#
+$
+%
+&
+'
+*
++
+-
+/
+0
+1
+2
+3
+4
+5
+6
+7
+8
+9
+:
+;
+<
+=
+>
+?
+A
+B
+C
+D
+E
+F
+G
+H
+I
+J
+K
+L
+M
+N
+O
+P
+Q
+R
+S
+T
+U
+V
+W
+X
+Y
+Z
+[
+\
+]
+^
+_
+`
+a
+b
+c
+d
+e
+f
+g
+h
+i
+j
+k
+l
+m
+n
+o
+p
+q
+r
+s
+t
+u
+v
+w
+x
+y
+z
+{
+|
+}
+~
+©
+°
+²
+½
+Á
+Ä
+Å
+Ç
+É
+Í
+Î
+Ó
+Ö
+×
+Ü
+ß
+à
+á
+â
+ã
+ä
+å
+æ
+ç
+è
+é
+ê
+ë
+ì
+í
+î
+ï
+ð
+ñ
+ò
+ó
+ô
+õ
+ö
+ø
+ú
+û
+ü
+ý
+ā
+ă
+ą
+ć
+Č
+č
+đ
+ē
+ė
+ę
+ě
+ğ
+ī
+İ
+ı
+Ł
+ł
+ń
+ň
+ō
+ř
+Ş
+ş
+Š
+š
+ţ
+ū
+ź
+ż
+Ž
+ž
+Ș
+ș
+Α
+Δ
+α
+λ
+φ
+Г
+О
+а
+в
+л
+о
+р
+с
+т
+я
+​
+’
+“
+”
+→
+∇
+∼
+「
+」
+ア
+カ
+グ
+ニ
+ラ
+ン
+ㄱ
+ㄴ
+ㄷ
+ㄸ
+ㄹ
+ㅂ
+ㅅ
+ㅆ
+ㅇ
+ㅈ
+ㅊ
+ㅋ
+ㅌ
+ㅎ
+ㅓ
+ㅜ
+ㅣ
+一
+丁
+七
+三
+上
+下
+不
+丑
+世
+丘
+丞
+中
+丸
+丹
+主
+乃
+久
+之
+乎
+乘
+九
+也
+乳
+乾
+事
+二
+云
+互
+五
+井
+亞
+亡
+交
+亥
+亨
+享
+京
+亭
+人
+仁
+今
+他
+仙
+代
+令
+以
+仰
+仲
+件
+任
+企
+伊
+伍
+伎
+伏
+伐
+休
+伯
+伴
+伸
+佃
+佈
+位
+低
+住
+佐
+何
+佛
+作
+使
+來
+供
+依
+侯
+侵
+侶
+便
+俗
+保
+俠
+信
+修
+俱
+俳
+倉
+個
+倍
+倒
+候
+借
+値
+倫
+倭
+假
+偈
+偉
+偏
+停
+偶
+傅
+傑
+傳
+傷
+傾
+像
+僞
+僥
+僧
+價
+儀
+儉
+儒
+優
+儼
+兀
+允
+元
+兆
+先
+光
+克
+兒
+入
+內
+全
+八
+公
+六
+共
+兵
+其
+具
+典
+兼
+再
+冠
+冥
+冶
+准
+凞
+凡
+凱
+出
+函
+刀
+分
+刊
+刑
+列
+初
+判
+別
+利
+到
+制
+券
+刺
+刻
+則
+前
+剛
+副
+創
+劃
+劑
+力
+功
+加
+劣
+助
+劫
+勇
+動
+務
+勝
+勢
+勳
+勸
+匈
+化
+北
+匠
+區
+十
+千
+午
+半
+卍
+卑
+卒
+卓
+南
+博
+卜
+占
+卦
+印
+危
+卵
+卷
+卽
+卿
+厄
+原
+厦
+去
+參
+又
+叉
+友
+反
+叔
+受
+口
+古
+句
+可
+台
+史
+右
+司
+各
+合
+吉
+同
+名
+后
+吏
+吐
+君
+吠
+吳
+呂
+告
+周
+味
+呵
+命
+和
+咳
+咸
+咽
+哀
+品
+哨
+哮
+哲
+唐
+唯
+唱
+商
+問
+啼
+善
+喆
+喉
+喜
+喩
+喪
+嘗
+器
+嚴
+囊
+四
+回
+因
+困
+固
+圈
+國
+圍
+園
+圓
+圖
+團
+土
+在
+地
+均
+坊
+坐
+坑
+坵
+型
+垢
+城
+域
+埴
+執
+培
+基
+堂
+堅
+堆
+堤
+堯
+報
+場
+塔
+塚
+塞
+塵
+境
+墜
+墟
+墨
+墳
+墾
+壁
+壇
+壓
+壤
+士
+壬
+壯
+壺
+壽
+夏
+夕
+外
+多
+夜
+夢
+大
+天
+太
+夫
+央
+失
+夷
+奄
+奇
+奉
+奎
+奏
+契
+奔
+奮
+女
+奴
+好
+如
+妄
+妊
+妖
+妙
+始
+姑
+姓
+姚
+姜
+威
+婆
+婚
+婦
+媒
+媚
+子
+孔
+字
+存
+孝
+孟
+季
+孤
+孫
+學
+孺
+宇
+守
+安
+宋
+宗
+官
+宙
+定
+客
+宣
+室
+宮
+害
+家
+容
+寂
+寃
+寄
+寅
+密
+寇
+富
+寒
+寓
+實
+審
+寫
+寬
+寶
+寸
+寺
+封
+將
+專
+尊
+對
+小
+少
+尙
+尹
+尼
+尿
+局
+居
+屈
+屋
+屍
+屎
+屛
+層
+屬
+山
+岐
+岡
+岩
+岳
+岸
+峙
+峰
+島
+峻
+峽
+崇
+崔
+崖
+崩
+嶋
+巖
+川
+州
+巢
+工
+左
+巧
+巨
+巫
+差
+己
+巷
+市
+布
+帝
+師
+帶
+常
+帽
+幕
+干
+平
+年
+幹
+幻
+幼
+幽
+庇
+序
+店
+府
+度
+座
+庫
+庭
+康
+廟
+廣
+廳
+延
+廷
+建
+廻
+弁
+式
+弑
+弓
+引
+弘
+弟
+弱
+張
+强
+弼
+彌
+彛
+形
+彬
+影
+役
+彼
+彿
+往
+征
+待
+律
+後
+徐
+徑
+得
+從
+循
+微
+德
+徹
+心
+必
+忌
+忍
+志
+忠
+思
+怡
+急
+性
+恐
+恒
+恨
+恩
+悅
+悖
+患
+悲
+情
+惑
+惟
+惠
+惡
+想
+惺
+愁
+意
+愚
+愛
+感
+愼
+慈
+態
+慕
+慣
+慧
+慾
+憂
+憤
+憺
+應
+懸
+戎
+成
+我
+戟
+戮
+戰
+戴
+戶
+房
+所
+手
+才
+打
+批
+承
+技
+抄
+把
+抗
+抱
+抽
+拇
+拓
+拘
+拙
+拜
+拾
+持
+指
+捌
+捨
+捿
+授
+掌
+排
+接
+推
+提
+揚
+揭
+援
+損
+搗
+摩
+播
+操
+擒
+擔
+擘
+據
+擧
+攘
+攝
+攬
+支
+改
+攻
+放
+政
+故
+敍
+敎
+救
+敗
+散
+敬
+整
+數
+文
+斗
+料
+斛
+斜
+斧
+斯
+新
+斷
+方
+於
+施
+旋
+族
+旗
+日
+旨
+早
+旱
+昌
+明
+易
+昔
+星
+春
+昧
+昭
+是
+時
+晉
+晋
+晩
+普
+景
+晴
+晶
+智
+暈
+暑
+暗
+暘
+曉
+曜
+曠
+曦
+曰
+曲
+書
+曹
+曼
+曾
+最
+會
+月
+有
+朋
+服
+望
+朝
+期
+木
+未
+末
+本
+朱
+朴
+李
+材
+村
+杖
+杜
+杞
+杭
+杯
+東
+松
+板
+林
+果
+枝
+枯
+枰
+枾
+柏
+柑
+柱
+栗
+校
+栢
+核
+根
+格
+桀
+桂
+案
+桎
+桑
+桓
+桔
+梁
+梏
+梓
+梗
+條
+梨
+梵
+棗
+棟
+森
+植
+椒
+楊
+楓
+楚
+業
+楮
+極
+榮
+槃
+槍
+樂
+樓
+樗
+樣
+樸
+樹
+樺
+樽
+橄
+橋
+橘
+機
+橡
+檀
+檎
+權
+欌
+欖
+次
+欲
+歌
+歐
+止
+正
+此
+步
+武
+歲
+歸
+死
+殖
+段
+殷
+殺
+殿
+毅
+母
+毒
+比
+毛
+氏
+民
+氣
+水
+永
+求
+汎
+汗
+江
+池
+沅
+沒
+沖
+沙
+沛
+河
+油
+治
+沼
+沿
+泉
+泊
+法
+泗
+泡
+波
+注
+泰
+洋
+洙
+洛
+洞
+津
+洲
+活
+派
+流
+浅
+浦
+浮
+浴
+海
+涅
+涇
+消
+涌
+液
+淑
+淡
+淨
+淫
+深
+淳
+淵
+淸
+渠
+渡
+游
+渾
+湖
+湯
+源
+溪
+溫
+溶
+滄
+滅
+滋
+滯
+滿
+漁
+漆
+漢
+漫
+漸
+潑
+潤
+潭
+澄
+澎
+澤
+澳
+澹
+濁
+濕
+濟
+濤
+濯
+瀋
+瀝
+灣
+火
+灰
+灸
+災
+炎
+炭
+点
+烈
+烏
+烙
+焚
+無
+焦
+然
+煌
+煎
+照
+煬
+煮
+熟
+熱
+燁
+燈
+燔
+燕
+燥
+燧
+燮
+爲
+爵
+父
+片
+版
+牌
+牛
+牝
+牟
+牡
+物
+特
+犧
+犬
+狀
+狗
+猥
+猩
+猪
+獨
+獵
+獸
+獻
+玄
+玉
+王
+玲
+珍
+珠
+珪
+班
+現
+球
+理
+琴
+瑞
+瑟
+瑪
+璃
+璋
+璽
+瓜
+瓦
+甑
+甘
+生
+産
+用
+甫
+田
+由
+甲
+申
+男
+界
+畏
+留
+畜
+畢
+略
+番
+異
+畵
+當
+畸
+疏
+疑
+疫
+疹
+疼
+病
+症
+痔
+痛
+痺
+瘀
+瘍
+瘡
+療
+癌
+癖
+登
+發
+白
+百
+的
+皆
+皇
+皮
+盂
+盆
+益
+盛
+盜
+盟
+盡
+盤
+盧
+目
+直
+相
+省
+看
+眞
+眼
+睡
+督
+瞋
+矢
+矣
+知
+短
+石
+破
+碍
+碑
+磁
+磨
+磬
+示
+社
+祇
+祖
+祝
+神
+祥
+祭
+祺
+禁
+禅
+禍
+福
+禦
+禪
+禮
+禹
+禽
+禾
+秀
+私
+秉
+秋
+科
+秘
+秤
+秦
+秩
+移
+稀
+稗
+種
+稱
+稷
+稼
+稽
+穀
+穆
+積
+空
+窮
+竅
+立
+章
+童
+竭
+端
+竹
+笑
+符
+第
+筆
+等
+筍
+答
+策
+箋
+箕
+管
+箱
+節
+篇
+簡
+米
+粉
+粘
+粥
+精
+糖
+糞
+系
+紀
+紂
+約
+紅
+紋
+純
+紙
+級
+素
+索
+紫
+紬
+累
+細
+紳
+終
+組
+結
+絡
+統
+絲
+絶
+絹
+經
+綠
+維
+綱
+網
+綸
+綽
+緖
+線
+緣
+緯
+縣
+縱
+總
+織
+繡
+繩
+繪
+繭
+纂
+續
+罕
+置
+罰
+羅
+羊
+美
+群
+義
+羽
+翁
+習
+翟
+老
+考
+者
+而
+耐
+耕
+耳
+聃
+聖
+聞
+聰
+聲
+職
+肇
+肉
+肖
+肝
+股
+肥
+育
+肺
+胃
+胎
+胚
+胞
+胡
+胥
+能
+脂
+脈
+脚
+脛
+脣
+脩
+脫
+脯
+脾
+腋
+腎
+腫
+腸
+腹
+膜
+膠
+膨
+膽
+臆
+臟
+臣
+臥
+臨
+自
+至
+致
+臺
+臼
+臾
+與
+興
+舊
+舌
+舍
+舒
+舜
+舟
+般
+船
+艦
+良
+色
+芋
+花
+芳
+芽
+苑
+苔
+苕
+苛
+苞
+若
+苦
+英
+茂
+茵
+茶
+茹
+荀
+荇
+草
+荒
+荷
+莊
+莫
+菊
+菌
+菜
+菩
+菫
+華
+菴
+菽
+萊
+萍
+萬
+落
+葉
+著
+葛
+董
+葬
+蒙
+蒜
+蒲
+蒸
+蒿
+蓮
+蔓
+蔘
+蔡
+蔬
+蕃
+蕉
+蕓
+薄
+薑
+薛
+薩
+薪
+薺
+藏
+藝
+藤
+藥
+藩
+藻
+蘆
+蘇
+蘊
+蘚
+蘭
+虎
+處
+虛
+虞
+虹
+蜀
+蜂
+蜜
+蝕
+蝶
+融
+蟬
+蟲
+蠶
+蠻
+血
+衆
+行
+術
+衛
+衡
+衣
+表
+袁
+裔
+裕
+裙
+補
+製
+複
+襄
+西
+要
+見
+視
+親
+覺
+觀
+角
+解
+言
+訂
+訊
+訓
+託
+記
+訣
+設
+診
+註
+評
+詩
+話
+詵
+誅
+誌
+認
+誕
+語
+誠
+誤
+誥
+誦
+說
+調
+談
+諍
+論
+諡
+諫
+諭
+諸
+謙
+講
+謝
+謠
+證
+識
+譚
+譜
+譯
+議
+護
+讀
+變
+谷
+豆
+豊
+豚
+象
+豪
+豫
+貝
+貞
+財
+貧
+貨
+貪
+貫
+貴
+貸
+費
+資
+賊
+賓
+賞
+賢
+賣
+賦
+質
+贍
+赤
+赫
+走
+起
+超
+越
+趙
+趣
+趨
+足
+趾
+跋
+跡
+路
+踏
+蹟
+身
+躬
+車
+軍
+軒
+軟
+載
+輓
+輕
+輪
+輯
+輸
+輻
+輿
+轅
+轉
+辨
+辭
+辯
+辰
+農
+近
+迦
+述
+追
+逆
+透
+逐
+通
+逝
+造
+逢
+連
+進
+逵
+遂
+遊
+運
+遍
+過
+道
+達
+遠
+遡
+適
+遷
+選
+遺
+遽
+還
+邊
+邑
+那
+邪
+郞
+郡
+部
+都
+鄒
+鄕
+鄭
+鄲
+配
+酒
+酸
+醉
+醫
+醯
+釋
+里
+重
+野
+量
+釐
+金
+針
+鈍
+鈴
+鉞
+銀
+銅
+銘
+鋼
+錄
+錢
+錦
+鎭
+鏡
+鐘
+鐵
+鑑
+鑛
+長
+門
+閃
+開
+間
+閔
+閣
+閥
+閭
+閻
+闕
+關
+阪
+防
+阿
+陀
+降
+限
+陝
+院
+陰
+陳
+陵
+陶
+陸
+陽
+隆
+隊
+隋
+階
+際
+障
+隣
+隨
+隱
+隷
+雀
+雄
+雅
+集
+雇
+雌
+雖
+雙
+雜
+離
+難
+雨
+雪
+雲
+電
+霜
+露
+靈
+靑
+靖
+靜
+非
+面
+革
+靴
+鞏
+韓
+音
+韶
+韻
+順
+須
+頊
+頌
+領
+頭
+顔
+願
+顚
+類
+顯
+風
+飛
+食
+飢
+飮
+飯
+飾
+養
+餓
+餘
+首
+香
+馨
+馬
+駒
+騫
+騷
+驕
+骨
+骸
+髓
+體
+高
+髥
+髮
+鬪
+鬱
+鬼
+魏
+魔
+魚
+魯
+鮮
+鰍
+鰐
+鳥
+鳧
+鳳
+鴨
+鵲
+鶴
+鷄
+鷹
+鹽
+鹿
+麗
+麥
+麻
+黃
+黑
+默
+點
+黨
+鼎
+齊
+齋
+齒
+龍
+龜
+가
+각
+간
+갇
+갈
+갉
+감
+갑
+값
+갓
+갔
+강
+갖
+갗
+같
+갚
+갛
+개
+객
+갠
+갤
+갬
+갭
+갯
+갰
+갱
+갸
+걀
+걔
+걘
+거
+걱
+건
+걷
+걸
+검
+겁
+것
+겄
+겅
+겆
+겉
+겊
+겋
+게
+겐
+겔
+겟
+겠
+겡
+겨
+격
+겪
+견
+결
+겸
+겹
+겻
+겼
+경
+곁
+계
+곕
+곗
+고
+곡
+곤
+곧
+골
+곪
+곬
+곯
+곰
+곱
+곳
+공
+곶
+과
+곽
+관
+괄
+괌
+광
+괘
+괜
+괭
+괴
+괸
+굉
+교
+구
+국
+군
+굳
+굴
+굵
+굶
+굼
+굽
+굿
+궁
+궂
+궈
+권
+궐
+궜
+궝
+궤
+귀
+귄
+귈
+귓
+규
+균
+귤
+그
+극
+근
+글
+긁
+금
+급
+긋
+긍
+기
+긴
+길
+김
+깁
+깃
+깅
+깊
+까
+깍
+깎
+깐
+깔
+깜
+깝
+깟
+깡
+깥
+깨
+깬
+깰
+깻
+깼
+깽
+꺄
+꺼
+꺽
+꺾
+껀
+껄
+껌
+껍
+껏
+껐
+껑
+께
+껴
+꼈
+꼍
+꼐
+꼬
+꼭
+꼴
+꼼
+꼽
+꼿
+꽁
+꽂
+꽃
+꽉
+꽝
+꽤
+꽥
+꾀
+꾜
+꾸
+꾹
+꾼
+꿀
+꿇
+꿈
+꿉
+꿋
+꿍
+꿎
+꿔
+꿨
+꿩
+꿰
+꿴
+뀄
+뀌
+뀐
+뀔
+뀜
+뀝
+끄
+끈
+끊
+끌
+끓
+끔
+끕
+끗
+끙
+끝
+끼
+끽
+낀
+낄
+낌
+낍
+낏
+낑
+나
+낙
+낚
+난
+낟
+날
+낡
+남
+납
+낫
+났
+낭
+낮
+낯
+낱
+낳
+내
+낵
+낸
+낼
+냄
+냅
+냇
+냈
+냉
+냐
+냔
+냘
+냥
+너
+넉
+넋
+넌
+널
+넓
+넘
+넙
+넛
+넜
+넝
+넣
+네
+넥
+넨
+넬
+넴
+넵
+넷
+넸
+넹
+녀
+녁
+년
+념
+녔
+녕
+녘
+녜
+노
+녹
+논
+놀
+놈
+놋
+농
+높
+놓
+놔
+놨
+뇌
+뇨
+뇩
+뇽
+누
+눅
+눈
+눌
+눔
+눕
+눗
+눠
+눴
+뉘
+뉜
+뉩
+뉴
+늄
+늅
+늉
+느
+늑
+는
+늘
+늙
+늠
+늡
+능
+늦
+늪
+늬
+니
+닉
+닌
+닐
+님
+닙
+닛
+닝
+닢
+다
+닥
+닦
+단
+닫
+달
+닭
+닮
+닯
+닳
+담
+답
+닷
+당
+닻
+닿
+대
+댁
+댄
+댈
+댐
+댑
+댓
+댔
+댕
+댜
+더
+덕
+덖
+던
+덜
+덟
+덤
+덥
+덧
+덩
+덫
+덮
+데
+덱
+덴
+델
+뎀
+뎃
+뎅
+뎌
+뎠
+뎨
+도
+독
+돈
+돋
+돌
+돔
+돕
+돗
+동
+돛
+돝
+돼
+됐
+되
+된
+될
+됨
+됩
+됴
+두
+둑
+둔
+둘
+둠
+둡
+둣
+둥
+둬
+뒀
+뒤
+뒬
+뒷
+뒹
+듀
+듈
+듐
+드
+득
+든
+듣
+들
+듦
+듬
+듭
+듯
+등
+듸
+디
+딕
+딘
+딛
+딜
+딤
+딥
+딧
+딨
+딩
+딪
+따
+딱
+딴
+딸
+땀
+땄
+땅
+때
+땐
+땔
+땜
+땝
+땠
+땡
+떠
+떡
+떤
+떨
+떫
+떰
+떱
+떳
+떴
+떵
+떻
+떼
+떽
+뗀
+뗄
+뗍
+뗏
+뗐
+뗑
+또
+똑
+똘
+똥
+뙤
+뚜
+뚝
+뚤
+뚫
+뚱
+뛰
+뛴
+뛸
+뜀
+뜁
+뜨
+뜩
+뜬
+뜯
+뜰
+뜸
+뜻
+띄
+띈
+띌
+띔
+띕
+띠
+띤
+띨
+띱
+띵
+라
+락
+란
+랄
+람
+랍
+랏
+랐
+랑
+랒
+랗
+래
+랙
+랜
+랠
+램
+랩
+랫
+랬
+랭
+랴
+략
+량
+러
+럭
+런
+럴
+럼
+럽
+럿
+렀
+렁
+렇
+레
+렉
+렌
+렐
+렘
+렙
+렛
+렝
+려
+력
+련
+렬
+렴
+렵
+렷
+렸
+령
+례
+로
+록
+론
+롤
+롬
+롭
+롯
+롱
+롸
+롹
+뢰
+뢴
+뢸
+룃
+료
+룐
+룡
+루
+룩
+룬
+룰
+룸
+룹
+룻
+룽
+뤄
+뤘
+뤼
+류
+륙
+륜
+률
+륨
+륭
+르
+륵
+른
+를
+름
+릅
+릇
+릉
+릎
+리
+릭
+린
+릴
+림
+립
+릿
+링
+마
+막
+만
+많
+맏
+말
+맑
+맘
+맙
+맛
+망
+맞
+맡
+맣
+매
+맥
+맨
+맬
+맴
+맵
+맷
+맸
+맹
+맺
+먀
+먁
+머
+먹
+먼
+멀
+멈
+멋
+멍
+멎
+메
+멕
+멘
+멜
+멤
+멥
+멧
+멩
+며
+멱
+면
+멸
+몄
+명
+몇
+모
+목
+몫
+몬
+몰
+몸
+몹
+못
+몽
+뫼
+묘
+무
+묵
+묶
+문
+묻
+물
+묽
+뭄
+뭅
+뭇
+뭉
+뭍
+뭏
+뭐
+뭔
+뭘
+뭡
+뭣
+뮈
+뮌
+뮐
+뮤
+뮬
+므
+믈
+믐
+미
+믹
+민
+믿
+밀
+밈
+밉
+밋
+밌
+밍
+및
+밑
+바
+박
+밖
+반
+받
+발
+밝
+밟
+밤
+밥
+밧
+방
+밭
+배
+백
+밴
+밸
+뱀
+뱁
+뱃
+뱄
+뱅
+뱉
+뱍
+뱐
+버
+벅
+번
+벌
+범
+법
+벗
+벙
+벚
+베
+벡
+벤
+벨
+벰
+벱
+벳
+벵
+벼
+벽
+변
+별
+볍
+볏
+볐
+병
+볕
+보
+복
+볶
+본
+볼
+봄
+봅
+봇
+봉
+봐
+봤
+뵈
+뵐
+뵙
+부
+북
+분
+붇
+불
+붉
+붐
+붓
+붕
+붙
+뷔
+뷰
+뷴
+뷸
+브
+븐
+블
+비
+빅
+빈
+빌
+빔
+빕
+빗
+빙
+빚
+빛
+빠
+빡
+빤
+빨
+빳
+빴
+빵
+빻
+빼
+빽
+뺀
+뺄
+뺌
+뺏
+뺐
+뺑
+뺨
+뻐
+뻑
+뻔
+뻗
+뻘
+뻣
+뻤
+뻥
+뻬
+뼈
+뼉
+뼘
+뽀
+뽈
+뽐
+뽑
+뽕
+뾰
+뿌
+뿍
+뿐
+뿔
+뿜
+쁘
+쁜
+쁠
+쁨
+삐
+삔
+삘
+사
+삭
+삯
+산
+살
+삵
+삶
+삼
+삽
+삿
+샀
+상
+샅
+새
+색
+샌
+샐
+샘
+샙
+샛
+샜
+생
+샤
+샨
+샬
+샴
+샵
+샷
+샹
+서
+석
+섞
+선
+섣
+설
+섬
+섭
+섯
+섰
+성
+섶
+세
+섹
+센
+셀
+셈
+셉
+셋
+셌
+셍
+셔
+션
+셜
+셨
+셰
+셴
+셸
+소
+속
+손
+솔
+솜
+솝
+솟
+송
+솥
+쇄
+쇠
+쇤
+쇳
+쇼
+숀
+숄
+숍
+수
+숙
+순
+숟
+술
+숨
+숩
+숫
+숭
+숯
+숱
+숲
+숴
+쉐
+쉘
+쉬
+쉭
+쉰
+쉴
+쉼
+쉽
+슈
+슐
+슘
+슛
+슝
+스
+슥
+슨
+슬
+슭
+슴
+습
+슷
+승
+시
+식
+신
+싣
+실
+싫
+심
+십
+싯
+싱
+싶
+싸
+싹
+싼
+쌀
+쌈
+쌉
+쌌
+쌍
+쌓
+쌔
+쌘
+쌩
+써
+썩
+썬
+썰
+썸
+썹
+썼
+썽
+쎄
+쎈
+쏘
+쏙
+쏜
+쏟
+쏠
+쏭
+쏴
+쐈
+쐐
+쐬
+쑤
+쑥
+쑨
+쒀
+쒔
+쓰
+쓱
+쓴
+쓸
+씀
+씁
+씌
+씨
+씩
+씬
+씰
+씸
+씹
+씻
+씽
+아
+악
+안
+앉
+않
+알
+앎
+앓
+암
+압
+앗
+았
+앙
+앞
+애
+액
+앤
+앨
+앰
+앱
+앳
+앴
+앵
+야
+약
+얀
+얄
+얇
+얌
+얍
+얏
+양
+얕
+얗
+얘
+얜
+어
+억
+언
+얹
+얻
+얼
+얽
+엄
+업
+없
+엇
+었
+엉
+엊
+엌
+엎
+에
+엑
+엔
+엘
+엠
+엡
+엣
+엥
+여
+역
+엮
+연
+열
+엷
+염
+엽
+엾
+엿
+였
+영
+옅
+옆
+옇
+예
+옌
+옐
+옙
+옛
+오
+옥
+온
+올
+옭
+옮
+옳
+옴
+옵
+옷
+옹
+옻
+와
+왁
+완
+왈
+왑
+왓
+왔
+왕
+왜
+왠
+왱
+외
+왼
+요
+욕
+욘
+욜
+욤
+용
+우
+욱
+운
+울
+움
+웁
+웃
+웅
+워
+웍
+원
+월
+웜
+웠
+웡
+웨
+웬
+웰
+웸
+웹
+위
+윅
+윈
+윌
+윔
+윗
+윙
+유
+육
+윤
+율
+윱
+윳
+융
+으
+윽
+은
+을
+읊
+음
+읍
+응
+의
+읜
+읠
+이
+익
+인
+일
+읽
+잃
+임
+입
+잇
+있
+잉
+잊
+잎
+자
+작
+잔
+잖
+잘
+잠
+잡
+잣
+잤
+장
+잦
+재
+잭
+잰
+잴
+잽
+잿
+쟀
+쟁
+쟈
+쟉
+쟤
+저
+적
+전
+절
+젊
+점
+접
+젓
+정
+젖
+제
+젝
+젠
+젤
+젬
+젭
+젯
+져
+젼
+졀
+졌
+졍
+조
+족
+존
+졸
+좀
+좁
+종
+좇
+좋
+좌
+좍
+좽
+죄
+죠
+죤
+주
+죽
+준
+줄
+줌
+줍
+줏
+중
+줘
+줬
+쥐
+쥔
+쥘
+쥬
+쥴
+즈
+즉
+즌
+즐
+즘
+즙
+증
+지
+직
+진
+짇
+질
+짊
+짐
+집
+짓
+징
+짖
+짙
+짚
+짜
+짝
+짠
+짢
+짤
+짧
+짬
+짭
+짰
+짱
+째
+짹
+짼
+쨀
+쨉
+쨋
+쨌
+쨍
+쩄
+쩌
+쩍
+쩐
+쩔
+쩜
+쩝
+쩡
+쩨
+쪄
+쪘
+쪼
+쪽
+쪾
+쫀
+쫄
+쫑
+쫓
+쫙
+쬐
+쭈
+쭉
+쭐
+쭙
+쯔
+쯤
+쯧
+찌
+찍
+찐
+찔
+찜
+찝
+찡
+찢
+찧
+차
+착
+찬
+찮
+찰
+참
+찹
+찻
+찼
+창
+찾
+채
+책
+챈
+챌
+챔
+챕
+챗
+챘
+챙
+챠
+챤
+처
+척
+천
+철
+첨
+첩
+첫
+청
+체
+첵
+첸
+첼
+쳄
+쳇
+쳉
+쳐
+쳔
+쳤
+초
+촉
+촌
+촘
+촛
+총
+촨
+촬
+최
+쵸
+추
+축
+춘
+출
+춤
+춥
+춧
+충
+춰
+췄
+췌
+취
+췬
+츄
+츠
+측
+츨
+츰
+층
+치
+칙
+친
+칠
+칡
+침
+칩
+칫
+칭
+카
+칵
+칸
+칼
+캄
+캅
+캇
+캉
+캐
+캔
+캘
+캠
+캡
+캣
+캤
+캥
+캬
+커
+컥
+컨
+컫
+컬
+컴
+컵
+컷
+컸
+컹
+케
+켄
+켈
+켐
+켓
+켕
+켜
+켠
+켤
+켭
+켯
+켰
+코
+콕
+콘
+콜
+콤
+콥
+콧
+콩
+콰
+콱
+콴
+콸
+쾅
+쾌
+쾡
+쾨
+쾰
+쿄
+쿠
+쿡
+쿤
+쿨
+쿰
+쿵
+쿼
+퀀
+퀄
+퀘
+퀭
+퀴
+퀵
+퀸
+퀼
+큐
+큘
+크
+큰
+클
+큼
+큽
+키
+킥
+킨
+킬
+킴
+킵
+킷
+킹
+타
+탁
+탄
+탈
+탉
+탐
+탑
+탓
+탔
+탕
+태
+택
+탠
+탤
+탬
+탭
+탯
+탰
+탱
+터
+턱
+턴
+털
+텀
+텁
+텃
+텄
+텅
+테
+텍
+텐
+텔
+템
+텝
+텡
+텨
+톈
+토
+톡
+톤
+톨
+톰
+톱
+톳
+통
+퇴
+툇
+투
+툭
+툰
+툴
+툼
+퉁
+퉈
+퉜
+튀
+튄
+튈
+튕
+튜
+튠
+튤
+튬
+트
+특
+튼
+튿
+틀
+틈
+틉
+틋
+틔
+티
+틱
+틴
+틸
+팀
+팁
+팅
+파
+팍
+팎
+판
+팔
+팜
+팝
+팟
+팠
+팡
+팥
+패
+팩
+팬
+팰
+팸
+팻
+팼
+팽
+퍼
+퍽
+펀
+펄
+펌
+펍
+펐
+펑
+페
+펙
+펜
+펠
+펨
+펩
+펫
+펭
+펴
+편
+펼
+폄
+폈
+평
+폐
+포
+폭
+폰
+폴
+폼
+폿
+퐁
+표
+푭
+푸
+푹
+푼
+풀
+품
+풋
+풍
+퓨
+퓬
+퓰
+퓸
+프
+픈
+플
+픔
+픕
+피
+픽
+핀
+필
+핌
+핍
+핏
+핑
+하
+학
+한
+할
+핥
+함
+합
+핫
+항
+해
+핵
+핸
+핼
+햄
+햅
+햇
+했
+행
+햐
+향
+헀
+허
+헉
+헌
+헐
+험
+헙
+헛
+헝
+헤
+헥
+헨
+헬
+헴
+헵
+헷
+헹
+혀
+혁
+현
+혈
+혐
+협
+혓
+혔
+형
+혜
+호
+혹
+혼
+홀
+홈
+홉
+홋
+홍
+홑
+화
+확
+환
+활
+홧
+황
+홰
+홱
+횃
+회
+획
+횝
+횟
+횡
+효
+후
+훅
+훈
+훌
+훑
+훔
+훗
+훤
+훨
+훼
+휄
+휑
+휘
+휙
+휜
+휠
+휩
+휭
+휴
+휼
+흄
+흉
+흐
+흑
+흔
+흘
+흙
+흠
+흡
+흣
+흥
+흩
+희
+흰
+흽
+히
+힉
+힌
+힐
+힘
+힙
+힝
+車
+滑
+金
+奈
+羅
+洛
+卵
+欄
+蘭
+郎
+來
+盧
+老
+魯
+綠
+鹿
+論
+雷
+樓
+縷
+凌
+樂
+不
+參
+葉
+沈
+若
+兩
+凉
+梁
+呂
+女
+廬
+麗
+黎
+曆
+歷
+戀
+蓮
+連
+列
+烈
+裂
+念
+獵
+靈
+領
+例
+禮
+醴
+惡
+尿
+料
+遼
+龍
+暈
+柳
+流
+類
+六
+陸
+倫
+律
+栗
+利
+李
+梨
+理
+離
+燐
+林
+臨
+立
+茶
+切
+宅
+ 

setup.py

@@ -32,7 +32,7 @@ setup(
     package_dir={'paddleocr': ''},
     include_package_data=True,
     entry_points={"console_scripts": ["paddleocr= paddleocr.paddleocr:main"]},
-    version='0.0.3',
+    version='2.0',
     install_requires=requirements,
     license='Apache License 2.0',
     description='Awesome OCR toolkits based on PaddlePaddle （8.6M ultra-lightweight pre-trained model, support training and deployment among server, mobile, embeded and IoT devices',

tools/export_model.py

@@ -39,26 +39,12 @@ def parse_args():
     return parser.parse_args()
 
 
-class Model(paddle.nn.Layer):
-    def __init__(self, model):
-        super(Model, self).__init__()
-        self.pre_model = model
-
-    # Please modify the 'shape' according to actual needs
-    @to_static(input_spec=[
-        paddle.static.InputSpec(
-            shape=[None, 3, 640, 640], dtype='float32')
-    ])
-    def forward(self, inputs):
-        x = self.pre_model(inputs)
-        return x
-
-
 def main():
     FLAGS = parse_args()
     config = load_config(FLAGS.config)
     logger = get_logger()
     # build post process
+
     post_process_class = build_post_process(config['PostProcess'],
                                             config['Global'])
 
@@ -71,9 +57,16 @@ def main():
     init_model(config, model, logger)
     model.eval()
 
-    model = Model(model)
-    save_path = '{}/{}'.format(FLAGS.output_path,
-                               config['Architecture']['model_type'])
+    save_path = '{}/{}/inference'.format(FLAGS.output_path,
+                                         config['Architecture']['model_type'])
+    infer_shape = [3, 32, 100] if config['Architecture'][
+        'model_type'] != "det" else [3, 640, 640]
+    model = to_static(
+        model,
+        input_spec=[
+            paddle.static.InputSpec(
+                shape=[None] + infer_shape, dtype='float32')
+        ])
     paddle.jit.save(model, save_path)
     logger.info('inference model is saved to {}'.format(save_path))
 

tools/infer/predict_cls.py

@@ -23,7 +23,7 @@ import copy
 import numpy as np
 import math
 import time
-
+import traceback
 import paddle.fluid as fluid
 
 import tools.infer.utility as utility
@@ -106,10 +106,10 @@ class TextClassifier(object):
                 norm_img_batch = fluid.core.PaddleTensor(norm_img_batch)
                 self.predictor.run([norm_img_batch])
             prob_out = self.output_tensors[0].copy_to_cpu()
-            cls_res = self.postprocess_op(prob_out)
+            cls_result = self.postprocess_op(prob_out)
             elapse += time.time() - starttime
-            for rno in range(len(cls_res)):
-                label, score = cls_res[rno]
+            for rno in range(len(cls_result)):
+                label, score = cls_result[rno]
                 cls_res[indices[beg_img_no + rno]] = [label, score]
                 if '180' in label and score > self.cls_thresh:
                     img_list[indices[beg_img_no + rno]] = cv2.rotate(
@@ -133,8 +133,8 @@ def main(args):
         img_list.append(img)
     try:
         img_list, cls_res, predict_time = text_classifier(img_list)
-    except Exception as e:
-        print(e)
+    except:
+        logger.info(traceback.format_exc())
         logger.info(
             "ERROR!!!! \n"
             "Please read the FAQ：https://github.com/PaddlePaddle/PaddleOCR#faq \n"
@@ -143,10 +143,10 @@ def main(args):
             "Please set --rec_image_shape='3,32,100' and --rec_char_type='en' ")
         exit()
     for ino in range(len(img_list)):
-        print("Predicts of {}:{}".format(valid_image_file_list[ino], cls_res[
+        logger.info("Predicts of {}:{}".format(valid_image_file_list[ino], cls_res[
             ino]))
-    print("Total predict time for {} images, cost: {:.3f}".format(
+    logger.info("Total predict time for {} images, cost: {:.3f}".format(
         len(img_list), predict_time))
 
-    if __name__ == "__main__":
-        main(utility.parse_args())
+if __name__ == "__main__":
+    main(utility.parse_args())

tools/infer/predict_det.py

@@ -178,11 +178,12 @@ if __name__ == "__main__":
         if count > 0:
             total_time += elapse
         count += 1
-        print("Predict time of {}: {}".format(image_file, elapse))
+        logger.info("Predict time of {}: {}".format(image_file, elapse))
         src_im = utility.draw_text_det_res(dt_boxes, image_file)
         img_name_pure = os.path.split(image_file)[-1]
         img_path = os.path.join(draw_img_save,
                                 "det_res_{}".format(img_name_pure))
         cv2.imwrite(img_path, src_im)
+        logger.info("The visualized image saved in {}".format(img_path))
     if count > 1:
-        print("Avg Time:", total_time / (count - 1))
+        logger.info("Avg Time:", total_time / (count - 1))

tools/infer/predict_rec.py

@@ -22,7 +22,7 @@ import cv2
 import numpy as np
 import math
 import time
-
+import traceback
 import paddle.fluid as fluid
 
 import tools.infer.utility as utility
@@ -135,8 +135,8 @@ def main(args):
         img_list.append(img)
     try:
         rec_res, predict_time = text_recognizer(img_list)
-    except Exception as e:
-        print(e)
+    except:
+        logger.info(traceback.format_exc())
         logger.info(
             "ERROR!!!! \n"
             "Please read the FAQ：https://github.com/PaddlePaddle/PaddleOCR#faq \n"
@@ -145,9 +145,9 @@ def main(args):
             "Please set --rec_image_shape='3,32,100' and --rec_char_type='en' ")
         exit()
     for ino in range(len(img_list)):
-        print("Predicts of {}:{}".format(valid_image_file_list[ino], rec_res[
+        logger.info("Predicts of {}:{}".format(valid_image_file_list[ino], rec_res[
             ino]))
-    print("Total predict time for {} images, cost: {:.3f}".format(
+    logger.info("Total predict time for {} images, cost: {:.3f}".format(
         len(img_list), predict_time))
 
 

tools/infer/predict_system.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 import os
 import sys
+
 __dir__ = os.path.dirname(os.path.abspath(__file__))
 sys.path.append(__dir__)
 sys.path.append(os.path.abspath(os.path.join(__dir__, '../..')))
@@ -23,17 +24,24 @@ import numpy as np
 import time
 from PIL import Image
 import tools.infer.utility as utility
-from tools.infer.utility import draw_ocr
 import tools.infer.predict_rec as predict_rec
 import tools.infer.predict_det as predict_det
+import tools.infer.predict_cls as predict_cls
 from ppocr.utils.utility import get_image_file_list, check_and_read_gif
 from ppocr.utils.logging import get_logger
+from tools.infer.utility import draw_ocr_box_txt
+
+logger = get_logger()
 
 
 class TextSystem(object):
     def __init__(self, args):
         self.text_detector = predict_det.TextDetector(args)
         self.text_recognizer = predict_rec.TextRecognizer(args)
+        self.use_angle_cls = args.use_angle_cls
+        self.drop_score = args.drop_score
+        if self.use_angle_cls:
+            self.text_classifier = predict_cls.TextClassifier(args)
 
     def get_rotate_crop_image(self, img, points):
         '''
@@ -72,12 +80,13 @@ class TextSystem(object):
         bbox_num = len(img_crop_list)
         for bno in range(bbox_num):
             cv2.imwrite("./output/img_crop_%d.jpg" % bno, img_crop_list[bno])
-            print(bno, rec_res[bno])
+            logger.info(bno, rec_res[bno])
 
     def __call__(self, img):
         ori_im = img.copy()
         dt_boxes, elapse = self.text_detector(img)
-        print("dt_boxes num : {}, elapse : {}".format(len(dt_boxes), elapse))
+        logger.info("dt_boxes num : {}, elapse : {}".format(
+            len(dt_boxes), elapse))
         if dt_boxes is None:
             return None, None
         img_crop_list = []
@@ -88,10 +97,23 @@ class TextSystem(object):
             tmp_box = copy.deepcopy(dt_boxes[bno])
             img_crop = self.get_rotate_crop_image(ori_im, tmp_box)
             img_crop_list.append(img_crop)
+        if self.use_angle_cls:
+            img_crop_list, angle_list, elapse = self.text_classifier(
+                img_crop_list)
+            logger.info("cls num  : {}, elapse : {}".format(
+                len(img_crop_list), elapse))
+
         rec_res, elapse = self.text_recognizer(img_crop_list)
-        print("rec_res num  : {}, elapse : {}".format(len(rec_res), elapse))
+        logger.info("rec_res num  : {}, elapse : {}".format(
+            len(rec_res), elapse))
         # self.print_draw_crop_rec_res(img_crop_list, rec_res)
-        return dt_boxes, rec_res
+        filter_boxes, filter_rec_res = [], []
+        for box, rec_reuslt in zip(dt_boxes, rec_res):
+            text, score = rec_reuslt
+            if score >= self.drop_score:
+                filter_boxes.append(box)
+                filter_rec_res.append(rec_reuslt)
+        return filter_boxes, filter_rec_res
 
 
 def sorted_boxes(dt_boxes):
@@ -107,8 +129,8 @@ def sorted_boxes(dt_boxes):
     _boxes = list(sorted_boxes)
 
     for i in range(num_boxes - 1):
-        if abs(_boxes[i+1][0][1] - _boxes[i][0][1]) < 10 and \
-            (_boxes[i + 1][0][0] < _boxes[i][0][0]):
+        if abs(_boxes[i + 1][0][1] - _boxes[i][0][1]) < 10 and \
+                (_boxes[i + 1][0][0] < _boxes[i][0][0]):
             tmp = _boxes[i]
             _boxes[i] = _boxes[i + 1]
             _boxes[i + 1] = tmp
@@ -119,7 +141,8 @@ def main(args):
     image_file_list = get_image_file_list(args.image_dir)
     text_sys = TextSystem(args)
     is_visualize = True
-    tackle_img_num = 0
+    font_path = args.vis_font_path
+    drop_score = args.drop_score
     for image_file in image_file_list:
         img, flag = check_and_read_gif(image_file)
         if not flag:
@@ -128,20 +151,12 @@ def main(args):
             logger.info("error in loading image:{}".format(image_file))
             continue
         starttime = time.time()
-        tackle_img_num += 1
-        if not args.use_gpu and args.enable_mkldnn and tackle_img_num % 30 == 0:
-            text_sys = TextSystem(args)
         dt_boxes, rec_res = text_sys(img)
         elapse = time.time() - starttime
-        print("Predict time of %s: %.3fs" % (image_file, elapse))
+        logger.info("Predict time of %s: %.3fs" % (image_file, elapse))
 
-        drop_score = 0.5
-        dt_num = len(dt_boxes)
-        for dno in range(dt_num):
-            text, score = rec_res[dno]
-            if score >= drop_score:
-                text_str = "%s, %.3f" % (text, score)
-                print(text_str)
+        for text, score in rec_res:
+            logger.info("{}, {:.3f}".format(text, score))
 
         if is_visualize:
             image = Image.fromarray(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))
@@ -149,18 +164,22 @@ def main(args):
             txts = [rec_res[i][0] for i in range(len(rec_res))]
             scores = [rec_res[i][1] for i in range(len(rec_res))]
 
-            draw_img = draw_ocr(
-                image, boxes, txts, scores, drop_score=drop_score)
+            draw_img = draw_ocr_box_txt(
+                image,
+                boxes,
+                txts,
+                scores,
+                drop_score=drop_score,
+                font_path=font_path)
             draw_img_save = "./inference_results/"
             if not os.path.exists(draw_img_save):
                 os.makedirs(draw_img_save)
             cv2.imwrite(
                 os.path.join(draw_img_save, os.path.basename(image_file)),
                 draw_img[:, :, ::-1])
-            print("The visualized image saved in {}".format(
+            logger.info("The visualized image saved in {}".format(
                 os.path.join(draw_img_save, os.path.basename(image_file))))
 
 
 if __name__ == "__main__":
-    logger = get_logger()
-    main(utility.parse_args())
+    main(utility.parse_args())
\ No newline at end of file

tools/infer/utility.py

@@ -71,6 +71,7 @@ def parse_args():
     parser.add_argument("--use_space_char", type=str2bool, default=True)
     parser.add_argument(
         "--vis_font_path", type=str, default="./doc/simfang.ttf")
+    parser.add_argument("--drop_score", type=float, default=0.5)
 
     # params for text classifier
     parser.add_argument("--use_angle_cls", type=str2bool, default=False)
@@ -99,8 +100,8 @@ def create_predictor(args, mode, logger):
     if model_dir is None:
         logger.info("not find {} model file path {}".format(mode, model_dir))
         sys.exit(0)
-    model_file_path = model_dir + "/model"
-    params_file_path = model_dir + "/params"
+    model_file_path = model_dir + ".pdmodel"
+    params_file_path = model_dir + ".pdiparams"
     if not os.path.exists(model_file_path):
         logger.info("not find model file path {}".format(model_file_path))
         sys.exit(0)
@@ -202,7 +203,12 @@ def draw_ocr(image,
     return image
 
 
-def draw_ocr_box_txt(image, boxes, txts):
+def draw_ocr_box_txt(image,
+                     boxes,
+                     txts,
+                     scores=None,
+                     drop_score=0.5,
+                     font_path="./doc/simfang.ttf"):
     h, w = image.height, image.width
     img_left = image.copy()
     img_right = Image.new('RGB', (w, h), (255, 255, 255))
@@ -212,7 +218,9 @@ def draw_ocr_box_txt(image, boxes, txts):
     random.seed(0)
     draw_left = ImageDraw.Draw(img_left)
     draw_right = ImageDraw.Draw(img_right)
-    for (box, txt) in zip(boxes, txts):
+    for idx, (box, txt) in enumerate(zip(boxes, txts)):
+        if scores is not None and scores[idx] < drop_score:
+            continue
         color = (random.randint(0, 255), random.randint(0, 255),
                  random.randint(0, 255))
         draw_left.polygon(box, fill=color)
@@ -228,8 +236,7 @@ def draw_ocr_box_txt(image, boxes, txts):
             1])**2)
         if box_height > 2 * box_width:
             font_size = max(int(box_width * 0.9), 10)
-            font = ImageFont.truetype(
-                "./doc/simfang.ttf", font_size, encoding="utf-8")
+            font = ImageFont.truetype(font_path, font_size, encoding="utf-8")
             cur_y = box[0][1]
             for c in txt:
                 char_size = font.getsize(c)
@@ -238,8 +245,7 @@ def draw_ocr_box_txt(image, boxes, txts):
                 cur_y += char_size[1]
         else:
             font_size = max(int(box_height * 0.8), 10)
-            font = ImageFont.truetype(
-                "./doc/simfang.ttf", font_size, encoding="utf-8")
+            font = ImageFont.truetype(font_path, font_size, encoding="utf-8")
             draw_right.text(
                 [box[0][0], box[0][1]], txt, fill=(0, 0, 0), font=font)
     img_left = Image.blend(image, img_left, 0.5)
@@ -254,7 +260,6 @@ def str_count(s):
     Count the number of Chinese characters,
     a single English character and a single number
     equal to half the length of Chinese characters.
-
     args:
         s(string): the input of string
     return(int):
@@ -289,7 +294,6 @@ def text_visual(texts,
         img_w(int): the width of blank img
         font_path: the path of font which is used to draw text
     return(array):
-
     """
     if scores is not None:
         assert len(texts) == len(