Summary of Triple-domain Feature Learning with Frequency-aware Memory Enhancement For Moving Infrared Small Target Detection, by Weiwei Duan et al.
Triple-domain Feature Learning with Frequency-aware Memory Enhancement for Moving Infrared Small Target Detection
by Weiwei Duan, Luping Ji, Shengjia Chen, Sicheng Zhu, Mao Ye
First submitted to arxiv on: 11 Jun 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a novel approach to moving infrared small target detection, which is challenging due to tiny targets and low contrast. The Triple-domain Strategy (Tridos) combines features from spatio-temporal, frequency, and memory domains to enhance feature representation. Tridos includes a local-global frequency-aware module that uses Fourier transform to detach and enhance frequency features, as well as a memory enhancement mechanism that captures spatial relations among video frames and encodes temporal dynamics via differential learning and residual enhancing. Additionally, the scheme incorporates residual compensation to reconcile cross-domain feature mismatches. Experimental results on three datasets (DAUB, ITSDT-15K, and IRDST) show that Tridos outperforms state-of-the-art methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Moving infrared small target detection is a challenging task due to tiny targets and low contrast. This paper proposes a new approach called Triple-domain Strategy (Tridos) that combines features from three domains: spatio-temporal, frequency, and memory. The goal is to enhance feature representation for better detection. The method uses Fourier transform to detach and enhance frequency features, captures spatial relations among video frames, and encodes temporal dynamics via differential learning and residual enhancing. |
