Denoising-Enhanced Coarse-to-Fine Infrared Small Target Detection with Attention Prior-Guided Knowledge Distillation

2026-06-20

AI Summary

This work addresses the challenge of detecting infrared small targets, which are difficult to identify due to their tiny size, weak features, and interference from complex dynamic backgrounds. Existing methods often suffer from redundant background computation and insufficient contextual utilization. To overcome these limitations, the authors propose ECFNet, a coarse-to-fine two-stage framework. In the coarse stage, a region binary classification network efficiently generates candidate regions, augmented by denoising auxiliary training to enhance discriminative capability. The fine stage integrates a lightweight detector with teacher–student cross-attention prior knowledge distillation to guide the model toward salient regions. By innovatively combining denoising auxiliary training and attention prior distillation, the proposed method achieves state-of-the-art performance on three real-world infrared datasets, outperforming both existing one-stage and two-stage approaches while maintaining high accuracy and real-time efficiency.

Abstract

Infrared small target detection (IRSTD) in high-resolution images is crucial for many practical applications, such as surveillance of unmanned aerial vehicles (UAVs) and UAV-based ground monitoring. However, IRSTD remains challenging due to the small size and weak features of targets, as well as significant interference from complex dynamic backgrounds. Existing detection methods often suffer from redundant computations on non-target background regions and insufficient exploitation of target context information, which limits their performance in complex backgrounds. To address these issues, we propose an efficient coarse-to-fine infrared small target detection framework with attention prior-guided knowledge distillation, termed ECFNet. In the coarse stage, we design a region binary classification network (RBCN) on grid-based multi-scale feature maps to efficiently recognize target-containing context region proposals while suppressing complex backgrounds. Moreover, we introduce a novel denoising-assisted training strategy that incorporates noisy ground-truth (GT) masks into the feature maps of RBCN and trains the network to reconstruct the GT masks through a denoising task, thereby enhancing its ability to distinguish target proposals from background regions and accelerating convergence. In the fine stage, we customize a lightweight target detector to the coarse stage's region proposals for balancing accuracy and efficiency. Furthermore, we propose a knowledge distillation strategy guided by the teacher-student cross-attention prior. This mechanism directs the student to focus on critical target regions, thereby enhancing the discriminative feature representation for infrared small targets. Extensive experiments on three real infrared datasets demonstrate that our method outperforms both existing single-stage and two-stage approaches while maintaining high real-time processing efficiency.

Problem

• Infrared small target detection
• complex dynamic backgrounds
• weak features
• target context information
• background interference

Innovation

• coarse-to-fine detection
• denoising-assisted training
• attention prior-guided knowledge distillation
• infrared small target detection
• region binary classification network