Diffuse to Detect: Bi-Level Sample Rebalancing with Pseudo-Label Diffusion for Point-Supervised Infrared Small-Target Detection

2026-05-20

AI Summary

This work addresses the challenges of unstable pseudo-label evolution and severe sample imbalance in pointly supervised infrared small target detection by proposing a dual-level joint optimization framework that integrates thermal radiation physical priors. The method employs a differentiable heat diffusion module to generate reliable pseudo-masks from single-point annotations and introduces a meta-classifier-driven dynamic sample reweighting mechanism to jointly optimize detector parameters, sample weights, and diffusion hyperparameters. Experimental results demonstrate that the proposed approach achieves performance comparable to fully supervised methods using only 30% of point-level annotations, yielding a fivefold improvement in annotation efficiency across multiple datasets while significantly outperforming existing state-of-the-art techniques in detection accuracy.

Abstract

Point supervision has become a scalable solution to address dense annotation for infrared small target detection, but its performance is limited by two coupled bottlenecks: unstable pseudo-label evolution in cluttered, low-contrast infrared imagery and severe sample-distribution imbalance. In this paper, we present a more adaptive and stable framework to address these issues. Leveraging the intrinsic consistency between thermal radiation patterns and heat diffusion, we propose a physics-induced annotation strategy that expands single-point labels into reliable pseudo-masks. To further enhance supervision and alleviate sample imbalance, we develop a bi-level dual-update framework that jointly optimizes detector weights, sample weights, and diffusion parameters. A meta-classifier dynamically predicts sample-wise loss weights, while a differentiable diffusion module refines pseudo-labels with detection feedback, enabling adaptive interaction between training and hyperparameter optimization. Extensive experiments across multiple datasets demonstrate five-fold annotation acceleration, superior detection accuracy, and comparable performance with 30% of the training data, validating the efficiency and practicality of our approach. Our code is available at https://github.com/yuanhang-yao/diffuse-to-detect.

Problem

• point-supervised learning
• infrared small target detection
• pseudo-label instability
• sample imbalance
• annotation efficiency

Innovation

• pseudo-label diffusion
• point-supervised learning
• sample rebalancing
• infrared small target detection
• bi-level optimization