Boosting Few-Shot Open-Set Object Detection via Prompt Learning and Robust Decision Boundary

TL;DR

This paper introduces a novel prompt-based framework for few-shot open-set object detection that leverages textual information and robust decision boundaries to improve unknown object rejection, outperforming previous methods.

Contribution

It proposes a new framework combining pseudo-unknown mining, evidence decoupling, and distribution calibration to enhance open-set detection with limited data.

Findings

Achieves 7.24% higher unknown recall on VOC10-5-5 dataset.

Improves unknown class detection by 1.38% on VOC-COCO dataset.

Outperforms previous state-of-the-art methods in open-set detection accuracy.

Abstract

Few-shot Open-set Object Detection (FOOD) poses a challenge in many open-world scenarios. It aims to train an open-set detector to detect known objects while rejecting unknowns with scarce training samples. Existing FOOD methods are subject to limited visual information, and often exhibit an ambiguous decision boundary between known and unknown classes. To address these limitations, we propose the first prompt-based few-shot open-set object detection framework, which exploits additional textual information and delves into constructing a robust decision boundary for unknown rejection. Specifically, as no available training data for unknown classes, we select pseudo-unknown samples with Attribution-Gradient based Pseudo-unknown Mining (AGPM), which leverages the discrepancy in attribution gradients to quantify uncertainty. Subsequently, we propose Conditional Evidence Decoupling (CED) to decouple and extract distinct knowledge from selected pseudo-unknown samples by eliminating opposing evidence. This optimization process can enhance the discrimination between known and unknown classes. To further regularize the model and form a robust decision boundary for unknown rejection, we introduce Abnormal Distribution Calibration (ADC) to calibrate the output probability distribution of local abnormal features in pseudo-unknown samples. Our method achieves superior performance over previous state-of-the-art approaches, improving the average recall of unknown class by 7.24% across all shots in VOC10-5-5 dataset settings and 1.38% in VOC-COCO dataset settings. Our source code is available at https://gitee.com/VR_NAVE/ced-food.