Overcoming Catastrophic Forgetting in Incremental Object Detection via Elastic Response Distillation

TL;DR

This paper introduces Elastic Response Distillation, a novel response-based incremental learning method for object detection that effectively mitigates catastrophic forgetting and improves performance on MS COCO.

Contribution

It proposes a new elastic response distillation technique focusing on response responses from classification and regression heads for incremental object detection.

Findings

Achieves state-of-the-art results on MS COCO

Significantly reduces performance gap with full training

Effectively retains localization and category knowledge

Abstract

Traditional object detectors are ill-equipped for incremental learning. However, fine-tuning directly on a well-trained detection model with only new data will lead to catastrophic forgetting. Knowledge distillation is a flexible way to mitigate catastrophic forgetting. In Incremental Object Detection (IOD), previous work mainly focuses on distilling for the combination of features and responses. However, they under-explore the information that contains in responses. In this paper, we propose a response-based incremental distillation method, dubbed Elastic Response Distillation (ERD), which focuses on elastically learning responses from the classification head and the regression head. Firstly, our method transfers category knowledge while equipping student detector with the ability to retain localization information during incremental learning. In addition, we further evaluate the quality of all locations and provide valuable responses by the Elastic Response Selection (ERS) strategy. Finally, we elucidate that the knowledge from different responses should be assigned with different importance during incremental distillation. Extensive experiments conducted on MS COCO demonstrate our method achieves state-of-the-art result, which substantially narrows the performance gap towards full training.