## Introduction
Multi Person PoseEstimation By PyTorch
## Results
<p align="left">
<img src="./readme/result.gif", width="720">
</p>
## Require
1. [Pytorch](http://pytorch.org/)
## Demo
- Download [converted pytorch model](https://www.dropbox.com/s/ae071mfm2qoyc8v/pose_model.pth?dl=0).
- Compile the C++ postprocessing: `cd lib/pafprocess; sh make.sh`
- `python demo/picture_demo.py` to run the picture demo.
- `python demo/web_demo.py` to run the web demo.
## Evalute
- `python evaluate/evaluation.py` to evaluate the model on coco val2017 dataset.
- It should have `mAP 0.653` for the rtpose, previous rtpose have `mAP 0.577` because we do left and right flip for heatmap and PAF for the evaluation.
c
### Main Results
| model name| mAP | Inference Time |
| :---------: | :---------: |:---------: |
|[original rtpose] | 0.653 |-|
Download link:
[rtpose](https://www.dropbox.com/s/ae071mfm2qoyc8v/pose_model.pth?dl=0)
## Development environment
The code is developed using python 3.6 on Ubuntu 18.04. NVIDIA GPUs are needed. The code is developed and tested using 4 1080ti GPU cards. Other platforms or GPU cards are not fully tested.
## Quick start
### 1. Preparation
#### 1.1 Prepare the dataset
- `cd training; bash getData.sh` to obtain the COCO 2017 images in `/data/root/coco/images/`, keypoints annotations in `/data/root/coco/annotations/`,
make them look like this:
```
${DATA_ROOT}
|-- coco
|-- annotations
|-- person_keypoints_train2017.json
|-- person_keypoints_val2017.json
|-- images
|-- train2017
|-- 000000000009.jpg
|-- 000000000025.jpg
|-- 000000000030.jpg
|-- ...
|-- val2017
|-- 000000000139.jpg
|-- 000000000285.jpg
|-- 000000000632.jpg
|-- ...
```
### 2. How to train the model
- Modify the data directory in `train/train_VGG19.py` and `python train/train_VGG19.py`
### Network Architecture
- testing architecture
![Teaser?](./readme/pose.png)
- training architecture
![Teaser?](./readme/training_structure.png)
没有合适的资源?快使用搜索试试~ 我知道了~
资源推荐
资源详情
资源评论
收起资源包目录
多人人体姿态估计_基于Pytorch实现的实时多人形体姿态估计算法_附项目源码_优质项目实战.zip (53个子文件)
多人人体姿态估计_基于Pytorch实现的实时多人形体姿态估计算法_附项目源码_优质项目实战
evaluate
evaluation.py 1KB
__init__.py 0B
coco_eval.py 10KB
lib
__init__.py 0B
utils
paf_to_pose.py 22KB
common.py 9KB
datasets
utils.py 1KB
__init__.py 47B
_init_paths.py 251B
coco.py 2KB
transforms.py 17KB
preprocessing.py 6KB
CocoDataDownloader.sh 541B
heatmap.py 1KB
datasets.py 13KB
paf.py 2KB
test_dataloader.py 10KB
network
__init__.py 0B
atrous_model.py 10KB
rtpose_mobilenetV2.py 4KB
post.py 23KB
atrous_model_share_stages.py 10KB
rtpose_hourglass.py 7KB
openpose.py 9KB
im_transform.py 4KB
atrouspose.py 6KB
rtpose_vgg.py 9KB
rtpose_shufflenetV2.py 9KB
pafprocess
__init__.py 0B
pafprocess.i 559B
setup.py 421B
numpy.i 107KB
pafprocess.cpp 9KB
pafprocess.h 1KB
README.md 319B
make.sh 67B
config
__init__.py 66B
default.py 4KB
train
_init_paths.py 251B
train_VGG19.py 13KB
train_ShuffleNetV2.py 11KB
train_SH.py 11KB
experiments
__init__.py 0B
vgg19_368x368_sgd.yaml 1KB
demo
picture_demo.py 2KB
web_demo.py 2KB
readme
training_structure.png 2.29MB
pose.png 2.16MB
ski.jpg 299KB
result.gif 50.05MB
requirements.txt 80B
video_demo.py 4KB
README.md 2KB
共 53 条
- 1
资源评论
__AtYou__
- 粉丝: 1644
- 资源: 507
下载权益
C知道特权
VIP文章
课程特权
开通VIP
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
最新资源
- 基于Java和深度学习的瓦斯浓度预测系统后端设计源码 - 瓦斯浓度预测后端
- Screenshot_20240528_103010.jpg
- 基于Python的新能源承载力计算及界面设计源码 - HAINING-DG
- 基于Java的本科探索学习项目设计源码 - 本科探索
- 基于Javascript和Python的微商城项目设计源码 - MicroMall
- 基于Java的网上订餐系统设计源码 - online ordering system
- 基于Javascript的超级美眉网络资源管理应用模块设计源码
- 基于Typescript和PHP的编程知识储备库设计源码 - study-php
- Screenshot_2024-05-28-11-40-58-177_com.tencent.mm.jpg
- 基于Dart的Flutter小提琴调音器APP设计源码 - violinhelper
资源上传下载、课程学习等过程中有任何疑问或建议,欢迎提出宝贵意见哦~我们会及时处理!
点击此处反馈
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功