LCPNet: Latent Consistent Proximal Unfolding Network for Infrared Small Target Detection
2026-07-06 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors focus on detecting tiny, distant objects in infrared images, which is tough because the background is complex. They note that current deep learning methods don't fully use the physical differences between targets and backgrounds, and existing improvements still mainly work in the original image space. Their new method, LCPNet, works in a 'latent' space where important features are clearer and uses a special solver and shared memory to better track and update target information. Experiments show their approach detects small targets more accurately and reliably than previous methods.
Infrared Small Target DetectionDeep LearningLatent SpaceLow-Rank PriorDeep UnfoldingProximal SolverShared Optimization MemoryRemote SensingImage-to-Mask Mapping
Authors
Tianfang Zhang, Fengyi Wu, Lei Li, Chang Liu, Zhenming Peng, Huaping Zhang, Xiangyang Ji
Abstract
Infrared small target detection (IRSTD) aims to identify long distance small targets from complex infrared backgrounds, and is a fundamental task in remote sensing. Deep learning methods have improved IRSTD by learning discriminative image-to-mask mappings, but such feed-forward designs often underuse physical decomposition structure between targets and backgrounds. Deep unfolding methods partially address this issue by embedding model-driven iterations into neural networks, yet existing designs still operate mainly in image domain and use updates and memory mechanisms that are not fully coupled with underlying optimization process. To address these limitations, we propose Latent Consistent Proximal unfolding network (LCPNet). First, we verify that low-rank prior remains valid in latent representations and perform unfolding in this space, preserving physical constraint while avoiding repeated compression of intermediate states. Second, we derive a Latent Consistent Proximal (LCP) solver that evolves each latent variable from its previous state rather than reconstructing through an indirect residual, and stabilizes small target updates through task-adaptive normalization and gain control. Third, we introduce Shared Optimization Memory (SOM), a common historical state shared by all decomposition variables to provide coordinated guidance across unfolding stages. Extensive experiments on four public benchmarks demonstrate that LCPNet outperforms state-of-the-art methods while achieving accurate and robust detection with low false alarms and competitive efficiency. Model and code are available at https://github.com/Tianfang-Zhang/LCPNet.