Skip to content

GitLab

Commit 2aa165a7 authored by wofmanaf's avatar wofmanaf Committed by zhe chen
Browse files

crowd_human detection

parent 76b26f09
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 1074 additions and 1 deletion
detection/configs/_base_/datasets/crowd_human.py 0 → 100644
View file @ 2aa165a7
# dataset settings
dataset_type = 'CrowdHumanDataset'
data_root = 'data/CrowdHuman/'
classes = ('person',)
img_norm_cfg = dict(
mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations', with_bbox=True),
dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
dict(type='RandomFlip', flip_ratio=0.5),
dict(type='Normalize', **img_norm_cfg),
dict(type='Pad', size_divisor=32),
dict(type='DefaultFormatBundle'),
dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(
type='MultiScaleFlipAug',
img_scale=(1333, 800),
flip=False,
transforms=[
dict(type='Resize', keep_ratio=True),
dict(type='RandomFlip'),
dict(type='Normalize', **img_norm_cfg),
dict(type='Pad', size_divisor=32),
dict(type='ImageToTensor', keys=['img']),
dict(type='Collect', keys=['img']),
])
]
data = dict(
samples_per_gpu=2,
workers_per_gpu=2,
train=dict(
type=dataset_type,
classes=classes,
filter_empty_gt=True,
ann_file=data_root + 'annotations/annotation_train.json',
img_prefix=data_root + 'Images',
pipeline=train_pipeline),
val=dict(
type=dataset_type,
classes=classes,
ann_file=data_root + 'annotations/annotation_val.json',
img_prefix=data_root + 'Images',
pipeline=test_pipeline),
test=dict(
type=dataset_type,
classes=classes,
ann_file=data_root + 'annotations/annotation_val.json',
img_prefix=data_root + 'Images',
pipeline=test_pipeline))
evaluation = dict(interval=100, metric='bbox')
detection/configs/_base_/models/cascade_mask_rcnn_r50_fpn_crowdhuman.py 0 → 100644
View file @ 2aa165a7
# model settings
model = dict(
type='CascadeRCNN',
backbone=dict(
type='ResNet',
depth=50,
num_stages=4,
out_indices=(0, 1, 2, 3),
frozen_stages=1,
norm_cfg=dict(type='BN', requires_grad=True),
norm_eval=True,
style='pytorch',
init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet50')),
neck=dict(
type='FPN',
in_channels=[256, 512, 1024, 2048],
out_channels=256,
num_outs=5),
rpn_head=dict(
type='RPNHead',
in_channels=256,
feat_channels=256,
anchor_generator=dict(
type='AnchorGenerator',
scales=[8],
ratios=[0.5, 1.0, 2.0],
strides=[4, 8, 16, 32, 64]),
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[.0, .0, .0, .0],
target_stds=[1.0, 1.0, 1.0, 1.0]),
loss_cls=dict(
type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
roi_head=dict(
type='CascadeRoIHead',
num_stages=3,
stage_loss_weights=[1, 0.5, 0.25],
bbox_roi_extractor=dict(
type='SingleRoIExtractor',
roi_layer=dict(type='RoIAlign', output_size=7, sampling_ratio=0),
out_channels=256,
featmap_strides=[4, 8, 16, 32]),
bbox_head=[
dict(
type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
num_classes=80,
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[0., 0., 0., 0.],
target_stds=[0.1, 0.1, 0.2, 0.2]),
reg_class_agnostic=True,
loss_cls=dict(
type='CrossEntropyLoss',
use_sigmoid=False,
loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0,
loss_weight=1.0)),
dict(
type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
num_classes=80,
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[0., 0., 0., 0.],
target_stds=[0.05, 0.05, 0.1, 0.1]),
reg_class_agnostic=True,
loss_cls=dict(
type='CrossEntropyLoss',
use_sigmoid=False,
loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0,
loss_weight=1.0)),
dict(
type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
num_classes=80,
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[0., 0., 0., 0.],
target_stds=[0.033, 0.033, 0.067, 0.067]),
reg_class_agnostic=True,
loss_cls=dict(
type='CrossEntropyLoss',
use_sigmoid=False,
loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
],),
# model training and testing settings
train_cfg=dict(
rpn=dict(
assigner=dict(
type='MaxIoUAssigner',
pos_iou_thr=0.7,
neg_iou_thr=0.3,
min_pos_iou=0.3,
match_low_quality=True,
ignore_iof_thr=-1),
sampler=dict(
type='RandomSampler',
num=256,
pos_fraction=0.5,
neg_pos_ub=-1,
add_gt_as_proposals=False),
allowed_border=0,
pos_weight=-1,
debug=False),
rpn_proposal=dict(
nms_pre=2000,
max_per_img=2000,
nms=dict(type='nms', iou_threshold=0.7),
min_bbox_size=0),
rcnn=[
dict(
assigner=dict(
type='MaxIoUAssigner',
pos_iou_thr=0.5,
neg_iou_thr=0.5,
min_pos_iou=0.5,
match_low_quality=False,
ignore_iof_thr=-1),
sampler=dict(
type='RandomSampler',
num=512,
pos_fraction=0.25,
neg_pos_ub=-1,
add_gt_as_proposals=True),
mask_size=28,
pos_weight=-1,
debug=False),
dict(
assigner=dict(
type='MaxIoUAssigner',
pos_iou_thr=0.6,
neg_iou_thr=0.6,
min_pos_iou=0.6,
match_low_quality=False,
ignore_iof_thr=-1),
sampler=dict(
type='RandomSampler',
num=512,
pos_fraction=0.25,
neg_pos_ub=-1,
add_gt_as_proposals=True),
mask_size=28,
pos_weight=-1,
debug=False),
dict(
assigner=dict(
type='MaxIoUAssigner',
pos_iou_thr=0.7,
neg_iou_thr=0.7,
min_pos_iou=0.7,
match_low_quality=False,
ignore_iof_thr=-1),
sampler=dict(
type='RandomSampler',
num=512,
pos_fraction=0.25,
neg_pos_ub=-1,
add_gt_as_proposals=True),
mask_size=28,
pos_weight=-1,
debug=False)
]),
test_cfg=dict(
rpn=dict(
nms_pre=1000,
max_per_img=1000,
nms=dict(type='nms', iou_threshold=0.7),
min_bbox_size=0),
rcnn=dict(
score_thr=0.05,
nms=dict(type='nms', iou_threshold=0.5),
max_per_img=100,
mask_thr_binary=0.5)))
detection/configs/crowd_human/README.md 0 → 100644
View file @ 2aa165a7
# CrowdHuman
## Introduction
Introduced by Shao et al. in [CrowdHuman: A Benchmark for Detecting Human in a Crowd](https://arxiv.org/pdf/1805.00123.pdf)
CrowdHuman is a benchmark dataset to better evaluate detectors in crowd scenarios. The CrowdHuman dataset is large, rich-annotated and contains high diversity. CrowdHuman contains 15000, 4370 and 5000 images for training, validation, and testing, respectively. There are a total of 470K human instances from train and validation subsets and 23 persons per image, with various kinds of occlusions in the dataset. Each human instance is annotated with a head bounding-box, human visible-region bounding-box and human full-body bounding-box. We hope our dataset will serve as a solid baseline and help promote future research in human detection tasks.
## Prepare the data
Download the original dataset from [CrowdHuman](https://www.crowdhuman.org/download.html). Then convert annotations by detection/tools/create_crowd_anno.py
- Data Tree of CrowdHuman should look like:
```bash
$ tree CrowdHuman
CrowdHuman
├── annotations
│ ├── annotation_train.json
│ ├── annotation_train.odgt
│ ├── annotation_val.json
│ ├── annotation_val.odgt
│ └── ...
└── Images
├── 1074488,79b360006b38332b.jpg
├── 1074488,79d54000c6f9d9e5.jpg
└── ...
## Model Zoo
### Cascade Mask R-CNN + InternImage
| backbone | schd | box mAP | mask mAP | train speed | train time | #param | FLOPs | Config | Download |
| :------------: | :---------: |:-------:|:--------:|:-----------:|:-----------:|:------:|:-----:| :---: |:--------:|
| InternImage-XL | 3x | TBD | TBD | TBD | TBD | TBD | TBD | [config](./cascade_internimage_xl_fpn_3x_crowd_human.py) | TBD |
- Training speed is measured with A100 GPUs using current code and may be faster than the speed in logs.
- Some logs are our recent newly trained ones. There might be slight differences between the results in logs and our paper.
detection/configs/crowd_human/cascade_internimage_xl_fpn_3x_crowd_human.py 0 → 100644
View file @ 2aa165a7
# --------------------------------------------------------
# InternImage
# Copyright (c) 2022 OpenGVLab
# Licensed under The MIT License [see LICENSE for details]
# --------------------------------------------------------
_base_ = [
'../_base_/models/cascade_mask_rcnn_r50_fpn_crowdhuman.py',
'../_base_/datasets/crowd_human.py',
'../_base_/schedules/schedule_3x.py',
'../_base_/default_runtime.py'
]
pretrained = 'https://github.com/OpenGVLab/InternImage/releases/download/cls_model/internimage_xl_22k_192to384.pth'
model = dict(
backbone=dict(
_delete_=True,
type='InternImage',
core_op='DCNv3',
channels=192,
depths=[5, 5, 24, 5],
groups=[12, 24, 48, 96],
mlp_ratio=4.,
drop_path_rate=0.4,
norm_layer='LN',
layer_scale=1.0,
offset_scale=2.0,
post_norm=True,
with_cp=False,
out_indices=(0, 1, 2, 3),
init_cfg=dict(type='Pretrained', checkpoint=pretrained)),
neck=dict(
type='FPN',
in_channels=[192, 384, 768, 1536],
out_channels=256,
num_outs=5),
rpn_head=dict(
type='RPNHead',
in_channels=256,
feat_channels=256,
anchor_generator=dict(
type='AnchorGenerator',
scales=[8],
ratios=[0.5, 1.0, 2.0],
strides=[4, 8, 16, 32, 64]),
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[.0, .0, .0, .0],
target_stds=[1.0, 1.0, 1.0, 1.0]),
loss_cls=dict(
type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
roi_head=dict(
bbox_head=[
dict(
type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
num_classes=1,
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[0.0, 0.0, 0.0, 0.0],
target_stds=[0.1, 0.1, 0.2, 0.2]),
reg_class_agnostic=True,
loss_cls=dict(
type='CrossEntropyLoss',
use_sigmoid=False,
loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0,
loss_weight=1.0)),
dict(
type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
num_classes=1,
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[0.0, 0.0, 0.0, 0.0],
target_stds=[0.05, 0.05, 0.1, 0.1]),
reg_class_agnostic=True,
loss_cls=dict(
type='CrossEntropyLoss',
use_sigmoid=False,
loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0,
loss_weight=1.0)),
dict(
type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
num_classes=1,
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[0.0, 0.0, 0.0, 0.0],
target_stds=[0.033, 0.033, 0.067, 0.067]),
reg_class_agnostic=True,
loss_cls=dict(
type='CrossEntropyLoss',
use_sigmoid=False,
loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
]))
img_norm_cfg = dict(
mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations', with_bbox=True),
dict(type='RandomFlip', flip_ratio=0.5),
dict(type='AutoAugment',
policies=[
[
dict(type='Resize',
img_scale=[(480, 1333), (512, 1333), (544, 1333), (576, 1333),
(608, 1333), (640, 1333), (672, 1333), (704, 1333),
(736, 1333), (768, 1333), (800, 1333)],
multiscale_mode='value',
keep_ratio=True)
],
[
dict(type='Resize',
img_scale=[(400, 1333), (500, 1333), (600, 1333)],
multiscale_mode='value',
keep_ratio=True),
dict(type='RandomCrop',
crop_type='absolute_range',
crop_size=(384, 600),
allow_negative_crop=True),
dict(type='Resize',
img_scale=[(480, 1333), (512, 1333), (544, 1333),
(576, 1333), (608, 1333), (640, 1333),
(672, 1333), (704, 1333), (736, 1333),
(768, 1333), (800, 1333)],
multiscale_mode='value',
override=True,
keep_ratio=True)
]
]),
dict(type='Normalize', **img_norm_cfg),
dict(type='Pad', size_divisor=32),
dict(type='DefaultFormatBundle'),
dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
]
# we use 8 nodes to train this model, with a total batch size of 128
data = dict(
samples_per_gpu=2,
train=dict(pipeline=train_pipeline))
optimizer = dict(
_delete_=True, type='AdamW', lr=0.0001, weight_decay=0.0001,
constructor='CustomLayerDecayOptimizerConstructor',
paramwise_cfg=dict(num_layers=39, layer_decay_rate=0.94,
depths=[5, 5, 24, 5]))
optimizer_config = dict(grad_clip=None)
# fp16 = dict(loss_scale=dict(init_scale=512))
evaluation = dict(save_best='auto')
checkpoint_config = dict(
interval=1,
max_keep_ckpts=3,
save_last=True,
)
detection/mmdet_custom/__init__.py
View file @ 2aa165a7
...@@ -4,4 +4,5 @@ ...@@ -4,4 +4,5 @@
# Licensed under The MIT License [see LICENSE for details] # Licensed under The MIT License [see LICENSE for details]
# -------------------------------------------------------- # --------------------------------------------------------
from .models import * # noqa: F401,F403 from .models import * # noqa: F401,F403
\ No newline at end of file from .datasets import *
\ No newline at end of file
detection/mmdet_custom/datasets/__init__.py 0 → 100644
View file @ 2aa165a7
# --------------------------------------------------------
# InternImage
# Copyright (c) 2022 OpenGVLab
# Licensed under The MIT License [see LICENSE for details]
# --------------------------------------------------------
from .crowd_human import CrowdHumanDataset
\ No newline at end of file
detection/mmdet_custom/datasets/crowd_human.py 0 → 100644
View file @ 2aa165a7
This diff is collapsed.
detection/tools/create_crowd_anno.py 0 → 100644
View file @ 2aa165a7
import argparse
import os
import pickle as pkl
import numpy as np
import random
from PIL import Image
import concurrent.futures
import json
import mmcv
def parse_args():
parser = argparse.ArgumentParser(description='Generate MMDetection Annotations for Crowdhuman-like dataset')
parser.add_argument('--dataset', help='dataset name', type=str)
parser.add_argument('--dataset-split', help='dataset split, e.g. train, val', type=str)
args = parser.parse_args()
return args.dataset, args.dataset_split
def load_func(fpath):
assert os.path.exists(fpath)
with open(fpath, 'r') as fid:
lines = fid.readlines()
records = [json.loads(line.strip('\n')) for line in lines]
return records
def decode_annotations(records, dataset_path):
rec_ids = list(range(len(records)))
img_list = []
ann_list = []
ann_id = 1
for idx, rec_id in enumerate(rec_ids):
img_id = records[rec_id]['ID']
img_url = dataset_path + 'Images/' + img_id + '.jpg'
assert os.path.exists(img_url)
im = Image.open(img_url)
im_w, im_h = im.width, im.height
gt_box = records[rec_id]['gtboxes']
gt_box_len = len(gt_box)
img_dict = dict(
file_name=img_id + '.jpg',
height=im_h,
width=im_w,
id=idx
)
img_list.append(img_dict)
for ii in range(gt_box_len):
each_data = gt_box[ii]
x, y, w, h = each_data['fbox']
if w <= 0 or h <= 0:
continue
# x1 = x; y1 = y; x2 = x + w; y2 = y + h
valid_bbox = [x, y, w, h]
if each_data['tag'] == 'person':
tag = 1
else:
tag = -2
if 'extra' in each_data:
if 'ignore' in each_data['extra']:
if each_data['extra']['ignore'] != 0:
tag = -2
ann_dict = dict(
area=w * h,
iscrowd=1 if tag == -2 else 0,
image_id=idx,
bbox=[x, y, w, h],
category_id=1,
id=ann_id,
# ignore=1 if tag == -2 else 1,
)
ann_id += 1
ann_list.append(ann_dict)
cate_list = [{'supercategory': 'none', 'id': 1, 'name': 'person'}]
json_dict = dict(
images=img_list,
annotations=ann_list,
categories=cate_list
)
return json_dict
if __name__ == "__main__":
dataset_name, dataset_type = parse_args()
dataset_path = 'data/%s/' % dataset_name
ch_file_path = dataset_path + 'annotations/annotation_%s.odgt' % dataset_type
json_file_path = dataset_path + 'annotations/annotation_%s.json' % dataset_type
records = load_func(ch_file_path)
print("Loading Annotations Done")
json_dict = decode_annotations(records, dataset_path)
print("Parsing Bbox Number: %d" % len(json_dict['annotations']))
mmcv.dump(json_dict, json_file_path)
detection/tools/evaluate/__init__.py 0 → 100644
View file @ 2aa165a7
from .compute_APMR import compute_APMR
from .compute_JI import compute_JI_with_ignore
\ No newline at end of file
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment