Deep learning improves resolution of clustered infrared targets
DISTA-Net++: Rethinking Infrared Small Target Unmixing Beyond Sub-Pixel Separation
Computer Vision and Pattern Recognition
Summary
Long-range infrared images often show small, closely packed heat sources as blurry blobs, making it hard to tell how many there are or exactly where they are. The authors argue that current methods, which try to pinpoint each source on a fixed grid, are too limited and can cause mistakes. They introduce a new way that uses counting information and continuous location adjustments to better separate these targets without needing super-fine grids. They also provide a large dataset and tools to help researchers test and improve these methods. Their approach shows more accurate counting and positioning of targets while using much less computing power.
What this means in practice
- •For defense imaging teams: Identify and count small heat sources in dense infrared scenes more accurately with less computing power for surveillance and reconnaissance.
- •For medical imaging developers: Improve detection of closely spaced small features in infrared-based medical scans by using continuous localization informed by overall object count.
Authors
Mengze Xu, Zhu Liu, Weidong Sheng, Boyang Li, Yimian Dai, Ming-Ming Cheng, Jian Yang
Abstract
Long-range infrared imaging frequently confronts dense target clusters whose diffraction-limited signatures merge into a single indistinguishable blob, concealing the number, sub-pixel positions, and radiant intensities of the underlying sources. While deep learning has advanced general object detection, resolving such Closely-Spaced Infrared Small Targets (CSIST) remains largely unexplored, owing to a systemic infrastructure void and a fundamental paradigm mismatch. The dominant formulation, which reduces unmixing to a blind, discrete sub-pixel separation, is inherently insufficient: without semantic guidance, the ill-posed inverse problem admits ambiguous solutions plagued by false and missed detections, while grid-based discretization locks predictions onto fixed lattice centers, chaining precision to prohibitively expensive grid refinement. We argue that CSIST unmixing should instead be informed and continuous. To ground this paradigm shift, we establish the first comprehensive open-source ecosystem for the field, comprising the large-scale CSIST-100K benchmark, a tailored metric suite, and the GrokCSO toolkit. Upon this foundation, we propose DISTA-Net++, which anchors a dynamic deep unfolding backbone with two synergistic mechanisms: a Count-Guided Prior that injects the global target count as an explicit semantic constraint to regularize the solution space, and a Continuous Coordinate Rectification that regresses off-grid offsets to decouple localization accuracy from grid resolution. Extensive experiments validate our paradigm: even under the most economical 3x division, DISTA-Net++ surpasses 7x-division state-of-the-art methods by 16.15% in CSO-mAP and 62.96% in count accuracy at merely one-sixth of their computation, demonstrating that unmixing precision need not be purchased with finer discretization. The complete ecosystem is available at https://github.com/GrokCV/GrokDet.