search

DirtyHarryLYL / HAKE-Action

As a part of the HAKE project, includes the reproduced SOTA models and the corresponding HAKE-enhanced versions (CVPR2020).
Apache License 2.0
100 stars 13 forks source link

怎么得到human part box #41

Closed whqwill closed 3 years ago

whqwill commented 3 years ago

作者你好，我看了一下Image-level-action的代码，目前训练还是测试都是直接用生成好的human box, object box还有human part box，我想问一下有没有代码去生成这样的box呢就是从原图

DirtyHarryLYL commented 3 years ago

Our script is here: https://github.com/DirtyHarryLYL/HAKE-Action/blob/Image-level-HAKE-Action/scripts/part_box_processing.py. You could first use AlphaPose to process all images and obtain the poses. Object detection can be finished via any detector, e.g., https://github.com/vt-vl-lab/iCAN#demotest-on-your-own-images.