Weichen Yu, Hongyuan Yu, Yan Huang, Liang Wang. "Generalized Inter-class Loss for Gait Recognition" in 30th ACM International Conference on Multimedia ( MM ’22) Main Track, October 10–14, 2022, Lisboa

In this paper, we propose we propose a generalized inter-class loss which resolves the inter-class variance from both sample-level feature distribution and class-level feature distribution. Instead of equal penalty strength on pair scores, the proposed loss optimizes sample-level inter-class feature distribution by dynamically adjusting the pairwise weight. Further, in class-level distribution, generalized inter-class loss adds a constraint on the uniformity of inter-class feature distribution, which forces the feature representations to approximate a hypersphere and keep maximal inter-class variance. In addition, the proposed method automatically adjusts the margin between classes which enables the inter-class feature distribution to be more flexible. The proposed method can be generalized to different gait recognition networks and achieves significant improvements.

1. Install Python 3.6, PyTorch 1.9.0.
2. Download data. You can obtain the dataset from [CASIA-B] or [OUMVLP].
3. Train the model run train.sh.

Python 3.6.5, Pytorch 1.1.0, Numpy 1.16.3, argparse and configparser

To replicate the results in CASIA-B and OUMVLP datasets in different backbones, use different configs in the "config" folder directly.

If you find this repo useful, please cite our paper.

@inproceedings{interclassloss,
  title     = {Generalized Inter-class Loss for Gait Recognition},
  author    = {Yu, Weichen and Yu, Hongyuan and Huang, Yan and Wang, Liang},
  booktitle = {Proceedings of the 30th ACM international conference on multimedia, {ACMMM-22}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  note      = {Main Track}
  doi       = {10.1145/3503161.3548311},
}

If you have any question or want to use the code, please contact yuweichen16@mails.ucas.ac.cn.

We appreciate the following Opengait github repo for their valuable code base:

https://github.com/ShiqiYu/OpenGait

We appreciate the support from Watrix.ai.