change name, delete data_path, change the order of the code

mmdet3d/core/evaluation/indoor_eval.py

@@ -4,6 +4,45 @@ import torch
 from mmdet3d.core.bbox.iou_calculators.iou3d_calculator import bbox_overlaps_3d
 
 
+def boxes3d_depth_to_lidar(boxes3d, mid_to_bottom=True):
+    """ Boxes3d Depth to Lidar.
+
+    Flip X-right,Y-forward,Z-up to X-forward,Y-left,Z-up.
+
+    Args:
+        boxes3d (ndarray): (N, 7) [x, y, z, w, l, h, r] in depth coords.
+
+    Return:
+        boxes3d_lidar (ndarray): (N, 7) [x, y, z, l, h, w, r] in LiDAR coords.
+    """
+    boxes3d_lidar = boxes3d.copy()
+    boxes3d_lidar[..., [0, 1, 2, 3, 4, 5]] = boxes3d_lidar[...,
+                                                           [1, 0, 2, 4, 3, 5]]
+    boxes3d_lidar[..., 1] *= -1
+    if mid_to_bottom:
+        boxes3d_lidar[..., 2] -= boxes3d_lidar[..., 5] / 2
+    return boxes3d_lidar
+
+
+def get_iou_gpu(bb1, bb2):
+    """Get IoU.
+
+    Compute IoU of two bounding boxes.
+
+    Args:
+        bb1 (ndarray): [x, y, z, w, l, h, ry] in LiDAR.
+        bb2 (ndarray): [x, y, z, h, w, l, ry] in LiDAR.
+
+    Returns:
+        ans_iou (tensor): The answer of IoU.
+    """
+
+    bb1 = torch.from_numpy(bb1).float().cuda()
+    bb2 = torch.from_numpy(bb2).float().cuda()
+    iou3d = bbox_overlaps_3d(bb1, bb2, mode='iou', coordinate='lidar')
+    return iou3d.cpu().numpy()
+
+
 def average_precision(recalls, precisions, mode='area'):
     """Calculate average precision (for single or multiple scales).
 
@@ -51,25 +90,6 @@ def average_precision(recalls, precisions, mode='area'):
     return ap
 
 
-def get_iou_gpu(bb1, bb2):
-    """Get IoU.
-
-    Compute IoU of two bounding boxes.
-
-    Args:
-        bb1 (ndarray): [x, y, z, w, l, h, ry] in LiDAR.
-        bb2 (ndarray): [x, y, z, h, w, l, ry] in LiDAR.
-
-    Returns:
-        ans_iou (tensor): The answer of IoU.
-    """
-
-    bb1 = torch.from_numpy(bb1).float().cuda()
-    bb2 = torch.from_numpy(bb2).float().cuda()
-    iou3d = bbox_overlaps_3d(bb1, bb2, mode='iou', coordinate='lidar')
-    return iou3d.cpu().numpy()
-
-
 def eval_det_cls(pred, gt, ovthresh=None):
     """Generic functions to compute precision/recall for object detection
         for a single class.
@@ -261,26 +281,6 @@ def eval_map_rec(det_infos, gt_infos, ovthresh=None):
     return rec, prec, ap
 
 
-def boxes3d_depth_to_lidar(boxes3d, mid_to_bottom=True):
-    """ Boxes3d Depth to Lidar.
-
-    Flip X-right,Y-forward,Z-up to X-forward,Y-left,Z-up.
-
-    Args:
-        boxes3d (ndarray): (N, 7) [x, y, z, w, l, h, r] in depth coords.
-
-    Return:
-        boxes3d_lidar (ndarray): (N, 7) [x, y, z, l, h, w, r] in LiDAR coords.
-    """
-    boxes3d_lidar = boxes3d.copy()
-    boxes3d_lidar[..., [0, 1, 2, 3, 4, 5]] = boxes3d_lidar[...,
-                                                           [1, 0, 2, 4, 3, 5]]
-    boxes3d_lidar[..., 1] *= -1
-    if mid_to_bottom:
-        boxes3d_lidar[..., 2] -= boxes3d_lidar[..., 5] / 2
-    return boxes3d_lidar
-
-
 def indoor_eval(gt_annos, dt_annos, metric, label2cat):
     """Scannet Evaluation.
 

mmdet3d/datasets/__init__.py

@@ -12,8 +12,8 @@ from .pipelines import (GlobalRotScale, IndoorFlipData, IndoorGlobalRotScale,
                         IndoorPointsColorNormalize, ObjectNoise,
                         ObjectRangeFilter, ObjectSample, PointShuffle,
                         PointsRangeFilter, RandomFlip3D)
-from .scannet_dataset import ScannetBaseDataset
-from .sunrgbd_dataset import SunrgbdBaseDataset
+from .scannet_dataset import ScanNetDataset
+from .sunrgbd_dataset import SUNRGBDDataset
 
 __all__ = [
     'KittiDataset', 'GroupSampler', 'DistributedGroupSampler',
@@ -23,6 +23,6 @@ __all__ = [
     'ObjectRangeFilter', 'PointsRangeFilter', 'Collect3D',
     'IndoorLoadPointsFromFile', 'IndoorPointsColorNormalize',
     'IndoorPointSample', 'IndoorLoadAnnotations3D', 'IndoorPointsColorJitter',
-    'IndoorGlobalRotScale', 'IndoorFlipData', 'SunrgbdBaseDataset',
-    'ScannetBaseDataset', 'IndoorBaseDataset'
+    'IndoorGlobalRotScale', 'IndoorFlipData', 'SUNRGBDDataset',
+    'ScanNetDataset', 'IndoorBaseDataset'
 ]

mmdet3d/datasets/indoor_base_dataset.py

@@ -32,6 +32,27 @@ class IndoorBaseDataset(torch_data.Dataset):
             self.pipeline = Compose(pipeline)
         self.with_label = with_label
 
+    def __len__(self):
+        return len(self.data_infos)
+
+    def get_data_info(self, index):
+        info = self.data_infos[index]
+        sample_idx = info['point_cloud']['lidar_idx']
+        pts_filename = self._get_pts_filename(sample_idx)
+
+        input_dict = dict(pts_filename=pts_filename)
+
+        if self.with_label:
+            annos = self._get_ann_info(index, sample_idx)
+            input_dict.update(annos)
+        if len(input_dict['gt_bboxes_3d']) == 0:
+            return None
+        return input_dict
+
+    def _rand_another(self, idx):
+        pool = np.where(self.flag == self.flag[idx])[0]
+        return np.random.choice(pool)
+
     def __getitem__(self, idx):
         if self.test_mode:
             return self.prepare_test_data(idx)
@@ -42,11 +63,6 @@ class IndoorBaseDataset(torch_data.Dataset):
                 continue
             return data
 
-    def prepare_test_data(self, index):
-        input_dict = self.get_data_info(index)
-        example = self.pipeline(input_dict)
-        return example
-
     def prepare_train_data(self, index):
         input_dict = self.get_data_info(index)
         if input_dict is None:
@@ -56,23 +72,10 @@ class IndoorBaseDataset(torch_data.Dataset):
             return None
         return example
 
-    def get_data_info(self, index):
-        info = self.data_infos[index]
-        sample_idx = info['point_cloud']['lidar_idx']
-        pts_filename = self._get_pts_filename(sample_idx)
-
-        input_dict = dict(pts_filename=pts_filename)
-
-        if self.with_label:
-            annos = self._get_ann_info(index, sample_idx)
-            input_dict.update(annos)
-        if len(input_dict['gt_bboxes_3d']) == 0:
-            return None
-        return input_dict
-
-    def _rand_another(self, idx):
-        pool = np.where(self.flag == self.flag[idx])[0]
-        return np.random.choice(pool)
+    def prepare_test_data(self, index):
+        input_dict = self.get_data_info(index)
+        example = self.pipeline(input_dict)
+        return example
 
     def _generate_annotations(self, output):
         """Generate Annotations.
@@ -126,6 +129,3 @@ class IndoorBaseDataset(torch_data.Dataset):
         gt_annos = [copy.deepcopy(info['annos']) for info in self.data_infos]
         ret_dict = indoor_eval(gt_annos, results, metric, self.label2cat)
         return ret_dict
-
-    def __len__(self):
-        return len(self.data_infos)

mmdet3d/datasets/scannet_dataset.py

@@ -7,7 +7,7 @@ from .indoor_base_dataset import IndoorBaseDataset
 
 
 @DATASETS.register_module()
-class ScannetBaseDataset(IndoorBaseDataset):
+class ScanNetDataset(IndoorBaseDataset):
 
     CLASSES = ('cabinet', 'bed', 'chair', 'sofa', 'table', 'door', 'window',
                'bookshelf', 'picture', 'counter', 'desk', 'curtain',
@@ -24,10 +24,8 @@ class ScannetBaseDataset(IndoorBaseDataset):
         super().__init__(root_path, ann_file, pipeline, classes, test_mode,
                          with_label)
 
-        self.data_path = osp.join(root_path, 'scannet_train_instance_data')
-
     def _get_pts_filename(self, sample_idx):
-        pts_filename = osp.join(self.data_path, f'{sample_idx}_vert.npy')
+        pts_filename = osp.join(self.root_path, f'{sample_idx}_vert.npy')
         return pts_filename
 
     def _get_ann_info(self, index, sample_idx):
@@ -41,9 +39,9 @@ class ScannetBaseDataset(IndoorBaseDataset):
             gt_bboxes_3d = np.zeros((1, 6), dtype=np.float32)
             gt_labels = np.zeros(1, ).astype(np.bool)
             gt_bboxes_3d_mask = np.zeros(1, ).astype(np.bool)
-        pts_instance_mask_path = osp.join(self.data_path,
+        pts_instance_mask_path = osp.join(self.root_path,
                                           f'{sample_idx}_ins_label.npy')
-        pts_semantic_mask_path = osp.join(self.data_path,
+        pts_semantic_mask_path = osp.join(self.root_path,
                                           f'{sample_idx}_sem_label.npy')
 
         anns_results = dict(

mmdet3d/datasets/sunrgbd_dataset.py

@@ -7,7 +7,7 @@ from .indoor_base_dataset import IndoorBaseDataset
 
 
 @DATASETS.register_module()
-class SunrgbdBaseDataset(IndoorBaseDataset):
+class SUNRGBDDataset(IndoorBaseDataset):
 
     CLASSES = ('bed', 'table', 'sofa', 'chair', 'toilet', 'desk', 'dresser',
                'night_stand', 'bookshelf', 'bathtub')
@@ -21,10 +21,9 @@ class SunrgbdBaseDataset(IndoorBaseDataset):
                  with_label=True):
         super().__init__(root_path, ann_file, pipeline, classes, test_mode,
                          with_label)
-        self.data_path = osp.join(root_path, 'sunrgbd_trainval')
 
     def _get_pts_filename(self, sample_idx):
-        pts_filename = osp.join(self.data_path, 'lidar',
+        pts_filename = osp.join(self.root_path, 'lidar',
                                 f'{sample_idx:06d}.npy')
         return pts_filename
 

tests/test_scannet_dataset.py

@@ -2,12 +2,12 @@ import numpy as np
 import pytest
 import torch
 
-from mmdet3d.datasets import ScannetBaseDataset
+from mmdet3d.datasets import ScanNetDataset
 
 
 def test_getitem():
     np.random.seed(0)
-    root_path = './tests/data/scannet'
+    root_path = './tests/data/scannet/scannet_train_instance_data'
     ann_file = './tests/data/scannet/scannet_infos.pkl'
     class_names = ('cabinet', 'bed', 'chair', 'sofa', 'table', 'door',
                    'window', 'bookshelf', 'picture', 'counter', 'desk',
@@ -36,7 +36,7 @@ def test_getitem():
             ]),
     ]
 
-    scannet_dataset = ScannetBaseDataset(root_path, ann_file, pipelines)
+    scannet_dataset = ScanNetDataset(root_path, ann_file, pipelines)
     data = scannet_dataset[0]
     points = data['points']._data
     gt_bboxes_3d = data['gt_bboxes_3d']._data
@@ -77,7 +77,7 @@ def test_evaluate():
         pytest.skip()
     root_path = './tests/data/scannet'
     ann_file = './tests/data/scannet/scannet_infos.pkl'
-    scannet_dataset = ScannetBaseDataset(root_path, ann_file)
+    scannet_dataset = ScanNetDataset(root_path, ann_file)
     results = []
     pred_boxes = dict()
     pred_boxes['box3d_lidar'] = np.array([[

tests/test_sunrgbd_dataset.py

@@ -2,12 +2,12 @@ import numpy as np
 import pytest
 import torch
 
-from mmdet3d.datasets import SunrgbdBaseDataset
+from mmdet3d.datasets import SUNRGBDDataset
 
 
 def test_getitem():
     np.random.seed(0)
-    root_path = './tests/data/sunrgbd'
+    root_path = './tests/data/sunrgbd/sunrgbd_trainval'
     ann_file = './tests/data/sunrgbd/sunrgbd_infos.pkl'
     class_names = ('bed', 'table', 'sofa', 'chair', 'toilet', 'desk',
                    'dresser', 'night_stand', 'bookshelf', 'bathtub')
@@ -28,7 +28,7 @@ def test_getitem():
         dict(type='Collect3D', keys=['points', 'gt_bboxes_3d', 'gt_labels']),
     ]
 
-    sunrgbd_dataset = SunrgbdBaseDataset(root_path, ann_file, pipelines)
+    sunrgbd_dataset = SUNRGBDDataset(root_path, ann_file, pipelines)
     data = sunrgbd_dataset[0]
     points = data['points']._data
     gt_bboxes_3d = data['gt_bboxes_3d']._data
@@ -67,7 +67,7 @@ def test_evaluate():
         pytest.skip()
     root_path = './tests/data/sunrgbd'
     ann_file = './tests/data/sunrgbd/sunrgbd_infos.pkl'
-    sunrgbd_dataset = SunrgbdBaseDataset(root_path, ann_file)
+    sunrgbd_dataset = SUNRGBDDataset(root_path, ann_file)
     results = []
     pred_boxes = dict()
     pred_boxes['box3d_lidar'] = np.array([[