Learning Muti-expert Distribution Calibration for Long-tailed Video Classification
This addresses the issue of imbalanced data in real-world video classification, which is a domain-specific problem, but the method is an incremental adaptation of long-tailed techniques from images to videos.
The paper tackles the problem of long-tailed class distribution in video classification, which causes model bias towards head classes and poor performance on tail classes, by proposing a multi-expert distribution calibration method that uses intra-class and inter-class distribution information to transfer knowledge from head to tail classes, achieving state-of-the-art results.
Most existing state-of-the-art video classification methods assume that the training data obey a uniform distribution. However, video data in the real world typically exhibit an imbalanced long-tailed class distribution, resulting in a model bias towards head class and relatively low performance on tail class. While the current long-tailed classification methods usually focus on image classification, adapting it to video data is not a trivial extension. We propose an end-to-end multi-expert distribution calibration method to address these challenges based on two-level distribution information. The method jointly considers the distribution of samples in each class (intra-class distribution) and the overall distribution of diverse data (inter-class distribution) to solve the issue of imbalanced data under long-tailed distribution. By modeling the two-level distribution information, the model can jointly consider the head classes and the tail classes and significantly transfer the knowledge from the head classes to improve the performance of the tail classes. Extensive experiments verify that our method achieves state-of-the-art performance on the long-tailed video classification task.