Weakly-supervised Contrastive Learning with Quantity Prompts for Moving Infrared Small Target Detection

2025-07-03

AI Summary

Mobile infrared small target detection faces challenges of high annotation costs, extremely low signal-to-clutter ratios, and sub-pixel-sized targets. Method: We propose WeCoL—the first weakly supervised contrastive learning framework for this task—requiring only video-level target count prompts for training. It innovatively integrates SAM-generated target activation maps, multi-frame energy accumulation, latent target mining, and long-/short-term motion-aware modeling, while establishing a contrastive learning mechanism in feature subspaces to enhance pseudo-label reliability. Contribution/Results: WeCoL is the first to introduce weak supervision into infrared small target detection, eliminating reliance on pixel-level annotations inherent in fully supervised paradigms. Experiments on DAUB and ITSDT-15K demonstrate that WeCoL achieves over 90% of the performance of state-of-the-art fully supervised methods, significantly reducing annotation overhead.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Computer Vision: Motion & Tracking
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Different from general object detection, moving infrared small target detection faces huge challenges due to tiny target size and weak background contrast.Currently, most existing methods are fully-supervised, heavily relying on a large number of manual target-wise annotations. However, manually annotating video sequences is often expensive and time-consuming, especially for low-quality infrared frame images. Inspired by general object detection, non-fully supervised strategies ($e.g.$, weakly supervised) are believed to be potential in reducing annotation requirements. To break through traditional fully-supervised frameworks, as the first exploration work, this paper proposes a new weakly-supervised contrastive learning (WeCoL) scheme, only requires simple target quantity prompts during model training.Specifically, in our scheme, based on the pretrained segment anything model (SAM), a potential target mining strategy is designed to integrate target activation maps and multi-frame energy accumulation.Besides, contrastive learning is adopted to further improve the reliability of pseudo-labels, by calculating the similarity between positive and negative samples in feature subspace.Moreover, we propose a long-short term motion-aware learning scheme to simultaneously model the local motion patterns and global motion trajectory of small targets.The extensive experiments on two public datasets (DAUB and ITSDT-15K) verify that our weakly-supervised scheme could often outperform early fully-supervised methods. Even, its performance could reach over 90% of state-of-the-art (SOTA) fully-supervised ones.

Problem

• Detect moving infrared small targets with weak supervision
• Reduce reliance on manual target-wise annotations
• Improve detection using contrastive learning and motion patterns

Innovation

• Weakly-supervised contrastive learning with quantity prompts
• Potential target mining using SAM and multi-frame energy
• Long-short term motion-aware learning for targets