Target Driven Adaptive Loss For Infrared Small Target Detection

2025-06-02

AI Summary

To address the limitations of conventional loss functions (e.g., BCE, IoU) in infrared small target detection (IRSTD)—specifically insufficient local region focus and poor robustness to small-scale and low-local-contrast targets—this paper proposes a target-driven adaptive loss function. Methodologically, it introduces: (1) a target-location-aware patch-wise weighting mechanism to strengthen supervision in target neighborhoods; and (2) a scale- and local-contrast-adaptive weight adjustment module that dynamically enhances gradient responses for challenging small targets. The proposed loss is architecture-agnostic and seamlessly integrates with mainstream semantic segmentation frameworks without requiring network modifications. Extensive experiments on three standard IRSTD benchmarks demonstrate consistent superiority over BCE, IoU, and other baselines across all key metrics: detection probability (Pd), false alarm rate (FAR), and F1-score—achieving state-of-the-art performance in each.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Local Search
• Machine Learning: Adversarial Learning & Robustness

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Security and Privacy: Large-scale security measurements

Abstract

We propose a target driven adaptive (TDA) loss to enhance the performance of infrared small target detection (IRSTD). Prior works have used loss functions, such as binary cross-entropy loss and IoU loss, to train segmentation models for IRSTD. Minimizing these loss functions guides models to extract pixel-level features or global image context. However, they have two issues: improving detection performance for local regions around the targets and enhancing robustness to small scale and low local contrast. To address these issues, the proposed TDA loss introduces a patch-based mechanism, and an adaptive adjustment strategy to scale and local contrast. The proposed TDA loss leads the model to focus on local regions around the targets and pay particular attention to targets with smaller scales and lower local contrast. We evaluate the proposed method on three datasets for IRSTD. The results demonstrate that the proposed TDA loss achieves better detection performance than existing losses on these datasets.

Problem

• Enhance infrared small target detection performance
• Improve robustness to small scale targets
• Address low local contrast issues

Innovation

• Target driven adaptive loss function
• Patch-based mechanism for local regions
• Adaptive adjustment for scale and contrast