项目作者： xmba15

项目描述：
Parsing Human Parts

高级语言： Jupyter Notebook

项目主页：

项目地址: git://github.com/xmba15/human_parsing_pytorch.git

创建时间： 2020-01-19T15:26:23Z
项目社区：https://github.com/xmba15/human_parsing_pytorch
开源协议：
下载

Pytorch Implementation of Human Parsing with PSPNet on LIP-Look Into Person-dataset

Dataset

Color Label Chart

Test Result

Steven Spielberg	Parsing Result

The approach in this github repository only utilizes pspnet for training a human parsing model, without the Self-supervised Structure-sensitive Learning skeme in [1]. Also, the training has been done with only a single gpu (with gradient accumulation). Therefore, the parsing result is not so good.

References

[1]. Look into Person: Self-supervised Structure-sensitive Learning and A New Benchmark for Human Parsing”, CVPR 2017

[2]. Pyramid Scene Parsing Network


