Dim and Small Target Detection with a Combined New Norm and Self-Attention Mechanism of Low-Rank Sparse Inversion

Area of Science:

• Computer Vision
• Signal Processing
• Remote Sensing

Background:

• Traditional infrared target detection methods struggle with clutter suppression and edge residuals in complex environments.
• Existing techniques often fail to effectively utilize the structural information within image data.

Purpose of the Study:

• To develop an advanced method for detecting small infrared targets in complex scenes.
• To overcome the limitations of traditional methods in clutter suppression and edge residual reduction.

Main Methods:

• Proposed a novel tensor nuclear norm based on linear transformation for global low-rank background constraint.
• Developed a self-attention mechanism to constrain target sparsity and eliminate edge residuals.
• Utilized the alternating direction multiplier method with a reweighted strategy for model optimization and accelerated convergence.

Main Results:

• Achieved average evaluation metric values of 0.9997 (SSIM), 467.23 (BSF), and 11.72 (SNR).
• Demonstrated superior detection rates compared to other algorithms at the same false alarm rate.
• Effectively suppressed background clutter and reduced edge residuals in detection results.

Conclusions:

• The proposed joint new norm and self-attention mechanism significantly improves infrared target detection in complex scenes.
• The method offers enhanced performance in terms of background suppression, edge residual reduction, and detection accuracy.
• This approach provides a robust solution for real-world applications requiring reliable small infrared target identification.