Curriculum Learning for Recurrent Video Object Segmentation

TL;DR

This paper investigates curriculum learning strategies, such as schedule sampling and frame skipping, to enhance recurrent video object segmentation, revealing that inverse schedule sampling and progressive frame skipping improve performance.

Contribution

It introduces novel curriculum learning techniques for recurrent video object segmentation, demonstrating their effectiveness over traditional methods.

Findings

Inverse schedule sampling outperforms forward sampling.

Progressive frame skipping benefits training with ground truth masks.

Improved performance on KITTI-MOTS car class.

Abstract

Video object segmentation can be understood as a sequence-to-sequence task that can benefit from the curriculum learning strategies for better and faster training of deep neural networks. This work explores different schedule sampling and frame skipping variations to significantly improve the performance of a recurrent architecture. Our results on the car class of the KITTI-MOTS challenge indicate that, surprisingly, an inverse schedule sampling is a better option than a classic forward one. Also, that a progressive skipping of frames during training is beneficial, but only when training with the ground truth masks instead of the predicted ones. Source code and trained models are available at http://imatge-upc.github.io/rvos-mots/.