Faster ILOD: Incremental Learning for Object Detectors based on Faster RCNN

TL;DR

This paper introduces a faster, more accurate incremental object detector based on Faster R-CNN that uses knowledge distillation to retain old knowledge and significantly improves efficiency over previous methods.

Contribution

It proposes an end-to-end incremental object detection method using multi-network adaptive distillation with Faster R-CNN, enhancing speed and accuracy.

Findings

Achieves 13 times faster detection than baseline methods.

Maintains high accuracy on PASCAL VOC and COCO datasets.

Effectively retains old category knowledge during incremental learning.

Abstract

The human vision and perception system is inherently incremental where new knowledge is continually learned over time whilst existing knowledge is retained. On the other hand, deep learning networks are ill-equipped for incremental learning. When a well-trained network is adapted to new categories, its performance on the old categories will dramatically degrade. To address this problem, incremental learning methods have been explored which preserve the old knowledge of deep learning models. However, the state-of-the-art incremental object detector employs an external fixed region proposal method that increases overall computation time and reduces accuracy comparing to Region Proposal Network (RPN) based object detectors such as Faster RCNN. The purpose of this paper is to design an efficient end-to-end incremental object detector using knowledge distillation. We first evaluate and analyze the performance of the RPN-based detector with classic distillation on incremental detection tasks. Then, we introduce multi-network adaptive distillation that properly retains knowledge from the old categories when fine-tuning the model for new task. Experiments on the benchmark datasets, PASCAL VOC and COCO, demonstrate that the proposed incremental detector based on Faster RCNN is more accurate as well as being 13 times faster than the baseline detector.