Disentangle Object and Non-object Infrared Features via Language Guidance

2026-01-14

AI Summary

This work addresses the challenge of infrared object detection, where low contrast and weak edges often result in insufficient discriminability of target features, particularly in complex scenes. To overcome this limitation, the study introduces— for the first time—a vision-language representation learning paradigm, proposing two key modules: Semantic Feature Alignment (SFA) and Object Feature Disentanglement (OFD). These modules leverage textual semantics to guide the disentanglement of target and non-target features, while incorporating correlation minimization to enhance the robustness of target representations. Evaluated on the M³FD and FLIR benchmarks, the proposed method achieves state-of-the-art performance with mAP scores of 83.7% and 86.1%, respectively, significantly outperforming existing approaches.

Technology Category

• Computer Vision: Language and Vision
• Machine Learning: Multimodal Learning
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

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.

Problem

• infrared object detection
• low contrast
• weak edge information
• feature disentanglement
• object detection

Innovation

• vision-language representation learning
• feature disentanglement
• infrared object detection
• semantic feature alignment
• object feature disentanglement