Probing Deep into Temporal Profile Makes the Infrared Small Target Detector Much Better

TL;DR

This work reframes infrared small target detection as a 1D temporal anomaly problem by leveraging long-term temporal saliency and target–background correlation in the temporal profile. It introduces a prediction-attribution tool to validate the primacy of temporal information, then presents the Temporal Probe (TPro) and a Deep Temporal Probe network (DeepPro) that perform all computations along the time dimension. DeepPro demonstrates state-of-the-art performance with substantially higher efficiency, particularly for ultra-dim targets and complex scenes, and achieves up to a 78.4% reduction in parameters compared to a lightweight baseline. Extending this idea, DeepPro-Plus combines temporal profiling with light spatial information, delivering further performance gains in noisy or real-world conditions, underscoring the practical impact of focusing on temporal-profile information for IRST detection.

Abstract

Infrared small target (IRST) detection is challenging in simultaneously achieving precise, robust, and efficient performance due to extremely dim targets and strong interference. Current learning-based methods attempt to leverage ``more" information from both the spatial and the short-term temporal domains, but suffer from unreliable performance under complex conditions while incurring computational redundancy. In this paper, we explore the ``more essential" information from a more crucial domain for the detection. Through theoretical analysis, we reveal that the global temporal saliency and correlation information in the temporal profile demonstrate significant superiority in distinguishing target signals from other signals. To investigate whether such superiority is preferentially leveraged by well-trained networks, we built the first prediction attribution tool in this field and verified the importance of the temporal profile information. Inspired by the above conclusions, we remodel the IRST detection task as a one-dimensional signal anomaly detection task, and propose an efficient deep temporal probe network (DeepPro) that only performs calculations in the time dimension for IRST detection. We conducted extensive experiments to fully validate the effectiveness of our method. The experimental results are exciting, as our DeepPro outperforms existing state-of-the-art IRST detection methods on widely-used benchmarks with extremely high efficiency, and achieves a significant improvement on dim targets and in complex scenarios. We provide a new modeling domain, a new insight, a new method, and a new performance, which can promote the development of IRST detection. Codes are available at https://github.com/TinaLRJ/DeepPro.

Figures (17)

• Figure 1: Visualizations of IRSTs in different scenarios and typical detection difficulties in different domains. (a, b) Distinct targets in different scenarios (Cases 1, 2, and 3): in appearance, targets and clutters show small inter-class differences, while different targets exhibit marked intra-class variations. (c1) Infrared image. (c2) Spatial domain: small dim targets are barely observable. (c3) Short-term spatial-temporal (ST) domain: targets are indistinguishable from interferences. (c4) Temporal profile: records dynamic statistical details of all signals at fixed spatial location, where small dim target signals are complete and more prominent. The temporal profile contains more essential information. (d, e) More contrasts of the three cases. For clarity, targets are marked with red annotations.
• Figure 2: Comparison of the detection performance, computational efficiency (Frames Per Second, FPS), and the number of parameters of different deep learning-based methods on dim targets. The color of each symbol represents the computational efficiency of the method. The yellower the color, the higher efficiency the method holds. Our proposed method is highly efficient and effective compared with other methods, including traditional multi-frame (MF) methods (SRSTT li2023sparse, 4DST-BTMD luo20234dst, and 4D-TR wu2023infrared), deep learning-based single-frame (SF) methods (DNA-Net li2022dense, UIUNet wu2022uiu, SCTransNet yuan2024sctransnet, and RPCANet wu2024rpcanet), and deep learning-based multi-frame methods (Res-UNet+DTUM li2023direction, STDMANet yan2023stdmanet, and Res-UNet+RFR ying2025infrared).
• Figure 3: Toy examples of the temporal profiles of different targets.
• Figure 4: Spatial and temporal-profile visualizations of targets in real scenes with noise and clutter. The red annotations mark the targets.
• Figure 5: Visualizations of the temporal profiles of target-in-noise(/-clutter) signals and correlation analyses between noise, clutter and target in the temporal profile. The target signals are same with fixed intensity (maximum is 1), and added to noise and clutter signals with different standard deviations (i.e., $\sigma_{N}$ and $\sigma_{B}$, respectively).