Disentangle Object and Non-object Infrared Features via Language Guidance

TL;DR

This paper introduces a novel vision-language learning framework for infrared object detection that uses textual guidance to disentangle object and non-object features, improving detection accuracy in challenging environments.

Contribution

It proposes a Semantic Feature Alignment and Object Feature Disentanglement modules that leverage textual supervision to enhance infrared object detection.

Findings

Achieves 83.7% mAP on M3FD benchmark.

Achieves 86.1% mAP on FLIR benchmark.

Outperforms existing methods in infrared object detection.

Abstract

Infrared object detection focuses on identifying and locating objects in complex environments (\eg, dark, snow, and rain) where visible imaging cameras are disabled by poor illumination. However, due to low contrast and weak edge information in infrared images, it is challenging to extract discriminative object features for robust detection. To deal with this issue, we propose a novel vision-language representation learning paradigm for infrared object detection. An additional textual supervision with rich semantic information is explored to guide the disentanglement of object and non-object features. Specifically, we propose a Semantic Feature Alignment (SFA) module to align the object features with the corresponding text features. Furthermore, we develop an Object Feature Disentanglement (OFD) module that disentangles text-aligned object features and non-object features by minimizing their correlation. Finally, the disentangled object features are entered into the detection head. In this manner, the detection performance can be remarkably enhanced via more discriminative and less noisy features. Extensive experimental results demonstrate that our approach achieves superior performance on two benchmarks: M\textsuperscript{3}FD (83.7\% mAP), FLIR (86.1\% mAP). Our code will be publicly available once the paper is accepted.