Multi-Task Incremental Learning for Object Detection

TL;DR

This paper introduces a multi-task incremental learning framework for object detection across various domains, addressing catastrophic forgetting with attentive feature distillation and adaptive exemplar sampling, achieving significant improvements.

Contribution

It proposes a novel incremental learning approach for object detection across multiple domains, incorporating attentive feature distillation and adaptive sampling to reduce forgetting.

Findings

Domain gaps have less negative impact than category differences.

Adaptive sampling effectively prevents forgetting in large domain and category shifts.

Significant performance improvements across seven benchmark datasets.

Abstract

Multi-task learns multiple tasks, while sharing knowledge and computation among them. However, it suffers from catastrophic forgetting of previous knowledge when learned incrementally without access to the old data. Most existing object detectors are domain-specific and static, while some are learned incrementally but only within a single domain. Training an object detector incrementally across various domains has rarely been explored. In this work, we propose three incremental learning scenarios across various domains and categories for object detection. To mitigate catastrophic forgetting, attentive feature distillation is proposed to leverages both bottom-up and top-down attentions to extract important information for distillation. We then systematically analyze the proposed distillation method in different scenarios. We find out that, contrary to common understanding, domain gaps have smaller negative impact on incremental detection, while category differences are problematic. For the difficult cases, where the domain gaps and especially category differences are large, we explore three different exemplar sampling methods and show the proposed adaptive sampling method is effective to select diverse and informative samples from entire datasets, to further prevent forgetting. Experimental results show that we achieve the significant improvement in three different scenarios across seven object detection benchmark datasets.