Breaking Immutable: Information-Coupled Prototype Elaboration for Few-Shot Object Detection

TL;DR

This paper introduces an innovative prototype elaboration method for few-shot object detection that adaptively generates query-specific prototypes, significantly improving detection accuracy by coupling query information and dynamically aggregating features.

Contribution

The paper proposes a novel Information-Coupled Prototype Elaboration (ICPE) method that enhances prototype quality by integrating query information and dynamically adjusting feature aggregation.

Findings

Achieves state-of-the-art results on Pascal VOC and MS COCO datasets.

Improves detection performance in few-shot settings across multiple experimental configurations.

Effectively highlights salient objects by adaptive feature aggregation.

Abstract

Few-shot object detection, expecting detectors to detect novel classes with a few instances, has made conspicuous progress. However, the prototypes extracted by existing meta-learning based methods still suffer from insufficient representative information and lack awareness of query images, which cannot be adaptively tailored to different query images. Firstly, only the support images are involved for extracting prototypes, resulting in scarce perceptual information of query images. Secondly, all pixels of all support images are treated equally when aggregating features into prototype vectors, thus the salient objects are overwhelmed by the cluttered background. In this paper, we propose an Information-Coupled Prototype Elaboration (ICPE) method to generate specific and representative prototypes for each query image. Concretely, a conditional information coupling module is introduced to couple information from the query branch to the support branch, strengthening the query-perceptual information in support features. Besides, we design a prototype dynamic aggregation module that dynamically adjusts intra-image and inter-image aggregation weights to highlight the salient information useful for detecting query images. Experimental results on both Pascal VOC and MS COCO demonstrate that our method achieves state-of-the-art performance in almost all settings.