Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MDAFNet: Multiscale Differential Edge and Adaptive Frequency Guided Network for Infrared Small Target Detection

Shuying Li, Qiang Ma, San Zhang, Wuwei Wang, Chuang Yang

TL;DR

Problem: infrared small target detection (IRSTD) suffers from cumulative edge degradation with network depth and insufficient frequency processing to separate high-frequency targets from high-frequency noise. Approach: MDAFNet integrates a Multi-Scale Differential Edge (MSDE) module and a Dual-Domain Adaptive Feature Enhancement (DAFE) module within a U-shaped encoder–decoder, leveraging Haar wavelet decomposition and adaptive strip pooling for frequency-aware enhancement. Contributions: MSDE preserves edge geometry across scales via differential edge operations and attention refinements, while DAFE performs adaptive frequency modulation to boost target signals and suppress noise; the method yields state-of-the-art performance on multiple IRSTD benchmarks, with improvements in $IoU$, $P_d$, and reductions in $F_a$. Significance: the framework improves detection reliability in cluttered infrared scenes and motivates future lightweight, real-time variants.

Abstract

Infrared small target detection (IRSTD) plays a crucial role in numerous military and civilian applications. However, existing methods often face the gradual degradation of target edge pixels as the number of network layers increases, and traditional convolution struggles to differentiate between frequency components during feature extraction, leading to low-frequency backgrounds interfering with high-frequency targets and high-frequency noise triggering false detections. To address these limitations, we propose MDAFNet (Multi-scale Differential Edge and Adaptive Frequency Guided Network for Infrared Small Target Detection), which integrates the Multi-Scale Differential Edge (MSDE) module and Dual-Domain Adaptive Feature Enhancement (DAFE) module. The MSDE module, through a multi-scale edge extraction and enhancement mechanism, effectively compensates for the cumulative loss of target edge information during downsampling. The DAFE module combines frequency domain processing mechanisms with simulated frequency decomposition and fusion mechanisms in the spatial domain to effectively improve the network's capability to adaptively enhance high-frequency targets and selectively suppress high-frequency noise. Experimental results on multiple datasets demonstrate the superior detection performance of MDAFNet.

MDAFNet: Multiscale Differential Edge and Adaptive Frequency Guided Network for Infrared Small Target Detection

TL;DR

Problem: infrared small target detection (IRSTD) suffers from cumulative edge degradation with network depth and insufficient frequency processing to separate high-frequency targets from high-frequency noise. Approach: MDAFNet integrates a Multi-Scale Differential Edge (MSDE) module and a Dual-Domain Adaptive Feature Enhancement (DAFE) module within a U-shaped encoder–decoder, leveraging Haar wavelet decomposition and adaptive strip pooling for frequency-aware enhancement. Contributions: MSDE preserves edge geometry across scales via differential edge operations and attention refinements, while DAFE performs adaptive frequency modulation to boost target signals and suppress noise; the method yields state-of-the-art performance on multiple IRSTD benchmarks, with improvements in , , and reductions in . Significance: the framework improves detection reliability in cluttered infrared scenes and motivates future lightweight, real-time variants.

Abstract

Infrared small target detection (IRSTD) plays a crucial role in numerous military and civilian applications. However, existing methods often face the gradual degradation of target edge pixels as the number of network layers increases, and traditional convolution struggles to differentiate between frequency components during feature extraction, leading to low-frequency backgrounds interfering with high-frequency targets and high-frequency noise triggering false detections. To address these limitations, we propose MDAFNet (Multi-scale Differential Edge and Adaptive Frequency Guided Network for Infrared Small Target Detection), which integrates the Multi-Scale Differential Edge (MSDE) module and Dual-Domain Adaptive Feature Enhancement (DAFE) module. The MSDE module, through a multi-scale edge extraction and enhancement mechanism, effectively compensates for the cumulative loss of target edge information during downsampling. The DAFE module combines frequency domain processing mechanisms with simulated frequency decomposition and fusion mechanisms in the spatial domain to effectively improve the network's capability to adaptively enhance high-frequency targets and selectively suppress high-frequency noise. Experimental results on multiple datasets demonstrate the superior detection performance of MDAFNet.
Paper Structure (16 sections, 6 equations, 5 figures, 5 tables)

This paper contains 16 sections, 6 equations, 5 figures, 5 tables.

Table of Contents

  1. Introduction
  2. METHOD
  3. Overall architecture
  4. Multi-Scale Differential Edge Module
  5. Dual-Domain Adaptive Feature Enhancement
  6. EXPERIMENTS
  7. Datasets and Implementation Setup
  8. Comparsion With SOTA Methods
  9. Quantitative Evaluations Results
  10. Qualitative Evaluations Results
  11. Ablation Study
  12. Conduct ablation on major components
  13. Selection of multi-scale branch numbers of MSDE
  14. Triple-path fusion effectiveness validation
  15. Ablation study on removing individual components from DAFE
  16. ...and 1 more sections

Figures (5)

  • Figure 1: Schematic diagram of the MDAFNet framework.
  • Figure 2: Architecture of the MSDE module.
  • Figure 3: Architecture of the DAFE module.
  • Figure 4: ROC curves comparison across three datasets.
  • Figure 5: Visual comparison of detection results for the proposed method against other mainstream approaches. Correctly detected targets are marked in blue, missed targets in yellow, and false alarms in red.