MAE-GEBD:Winning the CVPR'2023 LOVEU-GEBD Challenge

TL;DR

This paper presents an improved MAE-GEBD model for generic event boundary detection, achieving state-of-the-art results by refining data processing, loss functions, and segmentation strategies, and winning the CVPR 2023 LOVEU-GEBD challenge.

Contribution

The authors extend their previous MAE-GEBD approach with new data strategies, loss functions, and adaptive segmentation, leading to improved performance on the GEBD task.

Findings

Achieved 86.03% F1 score on Kinetics-GEBD test set.

Improved model performance with focal loss and dynamic segmentation.

Won the CVPR 2023 LOVEU-GEBD challenge.

Abstract

The Generic Event Boundary Detection (GEBD) task aims to build a model for segmenting videos into segments by detecting general event boundaries applicable to various classes. In this paper, based on last year's MAE-GEBD method, we have improved our model performance on the GEBD task by adjusting the data processing strategy and loss function. Based on last year's approach, we extended the application of pseudo-label to a larger dataset and made many experimental attempts. In addition, we applied focal loss to concentrate more on difficult samples and improved our model performance. Finally, we improved the segmentation alignment strategy used last year, and dynamically adjusted the segmentation alignment method according to the boundary density and duration of the video, so that our model can be more flexible and fully applicable in different situations. With our method, we achieve an F1 score of 86.03% on the Kinetics-GEBD test set, which is a 0.09% improvement in the F1 score compared to our 2022 Kinetics-GEBD method.