[简体中文](README_cn.md) | English
# PaddleVideo
## Introduction
![python version](https://img.shields.io/badge/python-3.7+-orange.svg) ![paddle version](https://img.shields.io/badge/PaddlePaddle-2.0-blue)
PaddleVideo is a toolset for video recognition, action localization, and spatio temporal action detection tasks prepared for the industry and academia. This repository provides examples and best practice guildelines for exploring deep learning algorithm in the scene of video area. We devote to support experiments and utilities which can significantly reduce the "time to deploy". By the way, this is also a proficiency verification and implementation of the newest PaddlePaddle 2.0 in the video field.
<div align="center">
<img src="docs/images/home.gif" width="450px"/><br>
</div>
### **If you think this repo is helpful to you, welcome to star us~ ⭐**
## Features
- **Various dataset and models**
PaddleVideo supports more datasets and models, including [Kinetics400](docs/zh-CN/dataset/k400.md), UCF101, YoutTube8M, NTU-RGB+D datasets, and video recognition models, such as TSN, TSM, SlowFast, TimeSformer, AttentionLSTM, ST-GCN and action localization model, like [BMN](./docs/zh-CN/model_zoo/localization/bmn.md).
- **Higher performance**
PaddleVideo has built-in solutions to improve accuracy on recognition models. [PP-TSM](docs/zh-CN/model_zoo/recognition/pp-tsm.md), which is based on the standard TSM, already archive the best performance in the 2D recognition network, has the same size of parameters but improve the Top1 Acc to 76.16%.
- **Faster training strategy**
PaddleVideo suppors faster training strategy, such as [AMP training](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/index_cn.html), Distributed training, Multigrid method for Slowfast, OP fusion method, Faster reader and so on.
- **Deployable**
PaddleVideo is powered by the Paddle Inference. There is no need to convert the model to ONNX format when deploying it, all you want can be found in this repository.
- **Applications**
PaddleVideo provides some interesting and practical projects that are implemented using video recognition and detection techniques, such as [FootballAction](https://github.com/PaddlePaddle/PaddleVideo/tree/application/FootballAction) and VideoTag.
### Overview of the performance
| Field | Model | Dataset | Metrics | ACC% |
| :--------------- | :--------: | :------------: | :------------: | :------------: |
| action recognition | [**PP-TSM**](./docs/zh-CN/model_zoo/recognition/pp-tsm.md) | [Kinetics-400](./docs/zh-CN/dataset/k400.md) | Top-1 | **76.16** |
| action recognition | [**PP-TSN**](./docs/zh-CN/model_zoo/recognition/pp-tsn.md) | [Kinetics-400](./docs/zh-CN/dataset/k400.md) | Top-1 | **75.06** |
| action recognition | [AGCN](./docs/zh-CN/model_zoo/recognition/agcn.md) | [FSD](./docs/zh-CN/dataset/fsd.md) | Top-1 | 90.66 |
| action recognition | [ST-GCN](./docs/zh-CN/model_zoo/recognition/stgcn.md) | [FSD](./docs/zh-CN/dataset/fsd.md) | Top-1 | 86.66 |
| action recognition | [TimeSformer](./docs/zh-CN/model_zoo/recognition/timesformer.md) | [Kinetics-400](./docs/zh-CN/dataset/k400.md) | Top-1 | 77.29 |
| action recognition | [SlowFast](./docs/zh-CN/model_zoo/recognition/slowfast.md) | [Kinetics-400](./docs/zh-CN/dataset/k400.md) | Top-1 | 75.84 |
| action recognition | [TSM](./docs/zh-CN/model_zoo/recognition/tsm.md) | [Kinetics-400](./docs/zh-CN/dataset/k400.md) | Top-1 | 71.06 |
| action recognition | [TSN](./docs/zh-CN/model_zoo/recognition/tsn.md) | [Kinetics-400](./docs/zh-CN/dataset/k400.md) | Top-1 | 69.81 |
| action recognition | [AttentionLSTM](./docs/zh-CN/model_zoo/recognition/attention_lstm.md) | [Youtube-8M](./docs/zh-CN/dataset/youtube8m.md) | Hit@1 | 89.0 |
| action detection| [BMN](./docs/zh-CN/model_zoo/localization/bmn.md) | [ActivityNet](./docs/zh-CN/dataset/ActivityNet.md) | AUC | 67.23 |
### Changelog
release/2.1 was released in 20/05/2021. Please refer to [release notes](https://github.com/PaddlePaddle/PaddleVideo/releases) for details.
<a name="Community"></a>
## Community
- Scan the QR code below with your Wechat and reply "video", you can access to official technical exchange group. Look forward to your participation.
<div align="center">
<img src="./docs/images/joinus.PNG" width = "200" height = "200" />
</div>
## Applications
- [VideoTag](https://github.com/PaddlePaddle/PaddleVideo/tree/application/VideoTag): 3k Large-Scale video classification model
<div align="center">
<img src="docs/images/VideoTag.gif" width="450px"/><br>
</div>
- [FootballAction](https://github.com/PaddlePaddle/PaddleVideo/tree/application/FootballAction): Football action detection model
<div align="center">
<img src="docs/images/FootballAction.gif" width="450px"/><br>
</div>
## Tutorials and Docs
- Tutorials and Slides
- [2021.01](https://aistudio.baidu.com/aistudio/course/introduce/6742)
- [Summarize of video understanding](docs/en/tutorials/summarize.md)
- Quick Start
- [Install](docs/en/install.md)
- [Start](docs/en/start.md)
- Project design
- [Modular design](docs/en/tutorials/modular_design.md)
- [Configuration design](docs/en/tutorials/config.md)
- Model zoo
- [recognition](docs/en/model_zoo/README.md)
- [TimeSformer](docs/en/model_zoo/recognition/timesformer.md)
- [Attention-LSTM](docs/en/model_zoo/recognition/attention_lstm.md)
- [TSN](docs/en/model_zoo/recognition/tsn.md)
- [TSM](docs/en/model_zoo/recognition/tsm.md)
- [PP-TSM](docs/en/model_zoo/recognition/pp-tsm.md)
- [PP-TSN](docs/en/model_zoo/recognition/pp-tsn.md)
- [SlowFast](docs/en/model_zoo/recognition/slowfast.md)
- [Localization](docs/en/model_zoo/README.md)
- [BMN](docs/en/model_zoo/localization/bmn.md)
- [Skeleton-based action recognition](docs/en/model_zoo/README.md)
- [ST-GCN](docs/en/model_zoo/recognition/stgcn.md)
- [AGCN](docs/en/model_zoo/recognition/agcn.md)
- Spatio temporal action detection
- Coming Soon!
- ActBERT: Learning Global-Local Video-Text Representations
- Coming Soon!
- Practice
- [Higher performance PP-TSM](docs/en/tutorials/pp-tsm.md)
- [Accelerate training](docs/en/tutorials/accelerate.md)
- [Deployment](docs/en/tutorials/deployment.md)
- Others
- [Benchmark](docs/en/benchmark.md)
- [Tools](docs/en/tools.md)
## License
PaddleVideo is released under the [Apache 2.0 license](LICENSE).
## Contributing
This poject welcomes contributions and suggestions. Please see our [contribution guidelines](docs/CONTRIBUTING.md).
- Many thanks to [mohui37](https://github.com/mohui37) for contributing the code for prediction.
没有合适的资源?快使用搜索试试~ 我知道了~
温馨提示
毕业设计基于骨骼点的花样滑冰动作识别python源码及说明.zip 【资源说明】 1.类别定义: 花样滑冰动作包括3个大类,分别为跳跃、旋转和步法,每个大类又包含很多小类。例如,跳跃大类包含:飞利浦三周跳(3Filp)和勾手三周跳(3Lutz)2个小类。然而,这2类跳跃的判别性仅在于一些个别帧的差异。此外,如果想就跳跃小类(3Filp或3Lutz)与旋转小类进行区别,对大部分帧的特征加以使用才能产生较好的判别性。 2.多义帧: 花样滑冰动作不同类别中相似的帧,甚至存在个别帧的特征相同等情况。 3.具体任务: 参赛选手利用比赛提供的训练集数据,构建基于骨骼点的细粒度动作识别模型,完成测试集的动作识别任务。模型识别效果由指标Accuracy排名决定,Accuracy得分越高,则认为该模型的动作识别效果越好。
资源推荐
资源详情
资源评论
收起资源包目录
毕业设计基于骨骼点的花样滑冰动作识别python源码及说明.zip (400个子文件)
Action Recognition Datasets 674B
Action Recognition Papers 2KB
example.avi 500KB
demos 318B
VideoTag.gif 10.05MB
home.gif 7.77MB
FootballAction.gif 5.72MB
.gitattributes 66B
MANIFEST.in 205B
acc_vps.jpeg 158KB
tsn_structure.jpg 491KB
slowfast_structure.jpg 476KB
i3d_expriment1.jpg 469KB
i3d_expand.jpg 459KB
i3d_expriment2.jpg 298KB
i3d_compare.jpg 289KB
slowfast_network.jpg 209KB
tsn_input.jpg 204KB
LICENSE 11KB
example_feat.list 66B
SlowFast.md 15KB
accelerate.md 12KB
summarize.md 12KB
README_cn.md 11KB
TSN.md 11KB
summarize.md 11KB
tsm.md 10KB
tsm.md 10KB
I3D.md 9KB
pp-tsn.md 8KB
pp-tsm.md 8KB
start.md 7KB
pp-tsn.md 7KB
项目必看.md 7KB
tsn.md 7KB
TSM.md 7KB
timesformer.md 7KB
README.md 7KB
pp-tsm.md 6KB
tsn.md 6KB
TSM.md 6KB
timesformer.md 6KB
whl_en.md 6KB
whl_zh.md 6KB
start.md 6KB
slowfast.md 6KB
README.md 6KB
bmn.md 5KB
README.md 5KB
slowfast.md 5KB
agcn.md 4KB
modular_design.md 4KB
stgcn.md 4KB
README.md 4KB
tsn_dali.md 4KB
config.md 4KB
agcn.md 4KB
ucf101.md 4KB
bmn.md 3KB
stgcn.md 3KB
config.md 3KB
AVA.md 3KB
ucf101.md 3KB
AVA.md 3KB
pp-tsm.md 3KB
k400.md 3KB
k400.md 3KB
tsn_dali.md 3KB
install.md 3KB
benchmark.md 2KB
fsd.md 2KB
benchmark.md 2KB
deployment.md 2KB
install.md 2KB
deployment.md 2KB
ppagcn.md 2KB
youtube8m.md 2KB
ntu-rgbd.md 1KB
attention_lstm.md 1KB
pp-tsm.md 1KB
ntu-rgbd.md 1KB
ActivityNet.md 1KB
youtube8m.md 998B
attention_lstm.md 835B
customized_usage.md 782B
ActivityNet.md 448B
tools.md 403B
tools.md 351B
CONTRIBUTING.md 300B
customized_usage.md 210B
modular_design.md 66B
accelerate.md 62B
example_skeleton.npy 1.43MB
example_feat.npy 313KB
BMN.png 853KB
action_framework.png 771KB
st-gcn.png 654KB
SlowFast.png 646KB
tsm_architecture.png 521KB
joinus.PNG 487KB
共 400 条
- 1
- 2
- 3
- 4
资源评论
- m0_626882382024-04-19资源内容详实,描述详尽,解决了我的问题,受益匪浅,学到了。
z同学的编程之路
- 粉丝: 1808
- 资源: 2129
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
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