Class-Incremental Learning for Action Recognition in Videos

Overview

This repository contains implementation for the paper "Class-Incremental Learning for Action Recognition in Videos" by Jaeyoo Park, Minsoo Kang and Bohyung Han, ICCV 2021.

We tackle catastrophic forgetting problem in the context of class-incremental learning for video recognition, which has not been explored actively despite the popularity of continual learning. Our framework addresses this challenging task by introducing time-channel importance maps and exploiting the importance maps for learning the representations of incoming examples via knowledge distillation. We also incorporate a regularization scheme in our objective function, which encourages individual features obtained from different time steps in a video to be uncorrelated and eventually improves accuracy by alleviating catastrophic forgetting. We evaluate the proposed approach on brand-new splits of class-incremental action recognition benchmarks constructed upon the UCF101, HMDB51, and Something-Something V2 datasets, and demonstrate the effectiveness of our algorithm in comparison to the existing continual learning methods that are originally designed for image data.

0.Installation

git clone https://github.com/bellos1203/TCD.git
cd TCD
pip install -r requirements.txt

1.Preprocessing for the videos

• We follow TSM codebase for the preprocessing.

2.Run the Model

• To reproduce the experimental results in the paper for UCF-101, please refer to the attached script files in scripts.
• Other experiments can be easily reproduced with slight modifications on .sh files in scripts or other options in opts.py.

3.Evaluation

• To obtain the numbers in the paper, compute average results from all 3 runs from 3 different seeds (1000, 1993, 2021).
• Note that we only run on single seed (1000) for Something-Something V2 datasets due to the computational cost.

Citation

If you use this code for your research, please cite our paper :

@inproceedings{park2021class,
title={Class-Incremental Learning for Action Recognition in Videos},
author={Park, Jaeyoo and Kang, Minsoo and Han, Bohyung},
booktitle={ICCV},
pages={13698--13707},
year={2021}
}

Comments

If you have any questions or comments, please contact me. bellos1203@snu.ac.kr

Acknowledgements

This code repository is basically developed based on TSM. Some parts related to incremental learning are borrowed from UCIR and PODNet.