AI实战营：MMPose开源算法库

guwuyue

已于 2023-06-05 06:53:26 修改

阅读量2.7k

点赞数 1

分类专栏：深度学习文章标签：人工智能深度学习计算机视觉

于 2023-06-04 00:20:31 首次发布

本文为博主原创文章，未经博主允许不得转载。

本文链接：https://blog.youkuaiyun.com/guwuyue/article/details/131025343

版权

深度学习专栏收录该内容

12 篇文章

订阅专栏

该文详细展示了如何运用MMPose库的预训练模型RTMPose和HRNet进行人体关键点检测，包括命令行预测单张图片和视频，以及使用官方的visualizer工具进行可视化。同时，文中还演示了通过PythonAPI进行模型预测和结果处理的步骤。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

目录

RTMPose关键点检测全流程

MMPose官方可视化工具visualizer

代码实现：
RTMPose关键点检测全流程
MMPose预训练模型预测-命令行
- 预测单张图
  - ```
  # HRNet
  python .\demo\topdown_demo_with_mmdet.py .\demo\mmdetection_cfg\faster_rcnn_r50_fpn_coco.py https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth .\configs\body_2d_keypoint\topdown_heatmap\coco\td-hm_hrnet-w32_8xb64-210e_coco-256x192.py https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w32_coco_256x192-c78dce93_20200708.pth --input .\data\test\multi-person.jpeg --output-root .\outputs\B1_HRNet_1 --dev cuda:0 --bbox-thr 0.5 --kpt-thr 0.2 --nms-thr 0.3 --radius 8 --thickness 4 --draw-bbox --draw-heatmap --show-kpt-idx
```
- Loads checkpoint by http backend from path: https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth
  Loads checkpoint by http backend from path: https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w32_coco_256x192-c78dce93_20200708.pth
  06/03 20:47:33 - mmengine - WARNING - `Visualizer` backend is not initialized because save_dir is None.
- ```
#RTMPose
python .\demo\topdown_demo_with_mmdet.py .\demo\mmdetection_cfg\faster_rcnn_r50_fpn_coco.py https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth .\projects\rtmpose\rtmpose\body_2d_keypoint\rtmpose-s_8xb256-420e_coco-256x192.py https://download.openmmlab.com/mmpose/v1/projects/rtmpose/rtmpose-s_simcc-aic-coco_pt-aic-coco_
420e-256x192-fcb2599b_20230126.pth --input .\data\test\multi-person.jpeg --output-root .\outputs\B1_RTM_1 --d
evice cuda:0 --bbox-thr 0.5 --kpt-thr 0.5 --nms-thr 0.3 --radius 8 --thickness 4 --draw-bbox --draw-heatmap --show-kpt-idx
```
  - Loads checkpoint by http backend from path: https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth
    Loads checkpoint by http backend from path: https://download.openmmlab.com/mmpose/v1/projects/rtmpose/rtmpose-s_simcc-aic-coco_pt-aic-coco_420e-256x192-fcb2599b_20230126.pth
    Downloading: "https://download.openmmlab.com/mmpose/v1/projects/rtmpose/rtmpose-s_simcc-aic-coco_pt-aic-coco_420e-256x192-fcb2599b_20230126.pth" to C:\Users\16798/.cache\torch\hub\checkpoints\rtmpose-s_simcc-aic-coco_pt-aic-coco_420e-256x192-fcb2599b_20230126.pth
    100%|██████████████████████████████████████████████████████████████████| 60.8M/60.8M [00:37<00:00, 1.71MB/s]
    06/03 23:04:00 - mmengine - WARNING - `Visualizer` backend is not initialized because save_dir is None.
    d:\workspace\.git\mmpose\mmpose\models\heads\coord_cls_heads\rtmcc_head.py:217: UserWarning: The predicted si
    mcc values are normalized for visualization. This may cause discrepancy between the keypoint scores and the 1D heatmaps.
    warnings.warn('The predicted simcc values are normalized for '
- 预测视频：直接将--input 换成视频路径即可
  - ```
  #HRNet 
  python .\demo\topdown_demo_with_mmdet.py .\demo\mmdetection_cfg\faster_rcnn_r50_fpn_coco.py https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth .\configs\body_2d_keypoint\topdown_heatmap\coco\td-hm_hrnet-w32_8xb64-210e_coco-256x192.py https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w32_coco_256x192-c78dce93_20200708.pth --input .\data\test\mother_wx.mp4 --output-root .\outputs\B1_HRNet_2 --device cuda:0 --bbox-thr 0.5 --kpt-thr 0.2 --nms-thr 0.3 --radius 5 --thickness 2 --draw-bbox --draw-heatmap --show-kpt-idx
```
- Loads checkpoint by http backend from path: https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth
  Loads checkpoint by http backend from path: https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w32_coco_256x192-c78dce93_20200708.pth
  06/03 23:11:03 - mmengine - WARNING - `Visualizer` backend is not initialized because save_dir is None.
- ```
#TRMPose
python .\demo\topdown_demo_with_mmdet.py .\demo\mmdetection_cfg\faster_rcnn_r50_fpn_coco.py https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth projects/rtmpose/rtmpose/body_2d_keypoint/rtmpose-s_8xb256-420e_coco-256x192.py https://download.openmmlab.com/mmpose/v1/projects/rtmpose/rtmpose-s_simcc-aic-coco_pt-aic-coco_420e-256x192-fcb2599b_20230126.pth --input data/test/mother_wx.mp4 --output-root outputs/B1_RTM_2 --device cuda:0 --bbox-thr 0.5 --kpt-thr 0.5 --nms-thr 0.3 --radius 5 --thickness 2 --draw-bbox --draw-heatmap  --show-kpt-idx
```
  - Loads checkpoint by http backend from path: https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth
    Loads checkpoint by http backend from path: https://download.openmmlab.com/mmpose/v1/projects/rtmpose/rtmpose-s_simcc-aic-coco_pt-aic-coco_420e-256x192-fcb2599b_20230126.pth
    06/03 23:14:39 - mmengine - WARNING - `Visualizer` backend is not initialized because save_dir is None.
    d:\workspace\.git\mmpose\mmpose\models\heads\coord_cls_heads\rtmcc_head.py:217: UserWarning: The predicted si
    mcc values are normalized for visualization. This may cause discrepancy between the keypoint scores and the 1D heatmaps.
    warnings.warn('The predicted simcc values are normalized for '

MMPose预测训练模型预测-Python API

import cv2
import numpy as np
from PIL import Image
import matplotlib.pyplot as plt
import torch
import mmcv
from mmcv import imread
import mmengine
from mmengine.registry import init_default_scope
from mmpose.apis import inference_topdown
from mmpose.apis import init_model as init_pose_estimator
from mmpose.evaluation.functional import nms
from mmpose.registry import VISUALIZERS
from mmpose.structures import merge_data_samples
from mmdet.apis import inference_detector, init_detector

# 设备
device = torch.device('cuda') if torch.cuda.is_available() else torch.device(
    'cpu')
print('device:', device)
# 载入待测图片
img_path = 'data/test/multi-person.jpeg'
# 打开图片
Image.open(img_path)
# 构建目标检测模型
# Faster RCNN
detector = init_detector(
    'demo/mmdetection_cfg/faster_rcnn_r50_fpn_coco.py',
    'https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth',
    device=device)
# RTMPose-Tiny
detector = init_detector(
    'projects/rtmpose/rtmdet/person/rtmdet_m_640-8xb32_coco-person.py',
    'https://download.openmmlab.com/mmpose/v1/projects/rtmpose/rtmdet_m_8xb32-100e_coco-obj365-person-235e8209.pth',
    device=device)
# 构建人体姿态估计模型
pose_estimator = init_pose_estimator(
    'configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_hrnet-w32_8xb64-210e_coco-256x192.py',
    'https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w32_coco_256x192-c78dce93_20200708.pth',
    device=device,
    cfg_options={'model': {
        'test_cfg': {
            'output_heatmaps': True
        }
    }})
# 预测-目标检测
init_default_scope(detector.cfg.get('default_scope', 'mmdet'))
#获取目标检测预测结果
detect_result = inference_detector(detector, img_path)
detect_result.keys()
# 预测类别
print(detect_result.pred_instances.labels)
# 置信度
print(detect_result.pred_instances.scores)
# 框坐标：左上角X坐标、左上角Y坐标、右下角X坐标、右下角Y坐标
print(detect_result.pred_instances.bboxes)
# 置信度阈值过滤，获得最终目标检测预测结果
# 置信度阈值
CONF_THRES = 0.5
pred_instance = detect_result.pred_instances.cpu().numpy()
bboxes = np.concatenate((pred_instance.bboxes, pred_instance.scores[:, None]), axis=1)
bboxes = bboxes[np.logical_and(pred_instance.labels == 0, pred_instance.scores > CONF_THRES)]
bboxes = bboxes[nms(bboxes, 0.3)][:, :4]
print(bboxes)
# 预测-关键点
# 获取每个 bbox 的关键点预测结果
pose_results = inference_topdown(pose_estimator, img_path, bboxes)
print(len(pose_results))
# 把多个bbox的pose结果打包到一起
data_samples = merge_data_samples(pose_results)
print(data_samples.keys())
# 预测结果-关键点坐标
print(data_samples.pred_instances.keypoints.shape)
# 索引为 0 的人，每个关键点的坐标
print(data_samples.pred_instances.keypoints[0,:,:])
# 预测结果-关键点热力图
# 每一类关键点的预测热力图
print(data_samples.pred_fields.heatmaps.shape)
idx_point = 13
heatmap = data_samples.pred_fields.heatmaps[idx_point,:,:]
print(heatmap.shape)
# 索引为 idx 的关键点，在全图上的预测热力图
plt.imshow(heatmap)
plt.show()

MMPose官方可视化工具`visualizer`

# 半径
pose_estimator.cfg.visualizer.radius = 10
# 线宽
pose_estimator.cfg.visualizer.line_width = 8
visualizer = VISUALIZERS.build(pose_estimator.cfg.visualizer)
# 元数据
visualizer.set_dataset_meta(pose_estimator.dataset_meta)
# 可视化
img = mmcv.imread(img_path)
img = mmcv.imconvert(img, 'bgr', 'rgb')

img_output = visualizer.add_datasample(
    'result',
    img,
    data_sample=data_samples,
    draw_gt=False,
    draw_heatmap=True,
    draw_bbox=True,
    show_kpt_idx=True,
    show=False,
    wait_time=0,
    out_file='outputs/B2.jpg')
print(img_output.shape)
plt.figure(figsize=(10, 10))
plt.imshow(img_output)
plt.show()