Generalized Few-Shot Object Detection without Forgetting

TL;DR

This paper introduces Retentive R-CNN, a novel few-shot object detection method that learns new classes without forgetting previous knowledge, achieving state-of-the-art results and maintaining performance on base classes.

Contribution

The paper proposes a simple yet effective approach combining Bias-Balanced RPN and Re-detector to prevent forgetting while learning new object classes in few-shot detection.

Findings

Retentive R-CNN outperforms existing methods on benchmarks.

It maintains base class performance while learning new classes.

Achieves competitive results on few-shot classes.

Abstract

Recently few-shot object detection is widely adopted to deal with data-limited situations. While most previous works merely focus on the performance on few-shot categories, we claim that detecting all classes is crucial as test samples may contain any instances in realistic applications, which requires the few-shot detector to learn new concepts without forgetting. Through analysis on transfer learning based methods, some neglected but beneficial properties are utilized to design a simple yet effective few-shot detector, Retentive R-CNN. It consists of Bias-Balanced RPN to debias the pretrained RPN and Re-detector to find few-shot class objects without forgetting previous knowledge. Extensive experiments on few-shot detection benchmarks show that Retentive R-CNN significantly outperforms state-of-the-art methods on overall performance among all settings as it can achieve competitive results on few-shot classes and does not degrade the base class performance at all. Our approach has demonstrated that the long desired never-forgetting learner is available in object detection.