Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MI-DETR: A Strong Baseline for Moving Infrared Small Target Detection with Bio-Inspired Motion Integration

Nian Liu, Jin Gao, Shubo Lin, Yutong Kou, Sikui Zhang, Fudong Ge, Zhiqiang Pu, Liang Li, Gang Wang, Yizheng Wang, Weiming Hu

TL;DR

A bio-inspired dual-pathway detector that processes one infrared frame per time step while explicitly modeling motion while explicitly modeling motion is proposed, demonstrating the effectiveness of biologically inspired motion-appearance integration.

Abstract

Infrared small target detection (ISTD) is challenging because tiny, low-contrast targets are easily obscured by complex and dynamic backgrounds. Conventional multi-frame approaches typically learn motion implicitly through deep neural networks, often requiring additional motion supervision or explicit alignment modules. We propose Motion Integration DETR (MI-DETR), a bio-inspired dual-pathway detector that processes one infrared frame per time step while explicitly modeling motion. First, a retina-inspired cellular automaton (RCA) converts raw frame sequences into a motion map defined on the same pixel grid as the appearance image, enabling parvocellular-like appearance and magnocellular-like motion pathways to be supervised by a single set of bounding boxes without extra motion labels or alignment operations. Second, a Parvocellular-Magnocellular Interconnection (PMI) Block facilitates bidirectional feature interaction between the two pathways, providing a biologically motivated intermediate interconnection mechanism. Finally, a RT-DETR decoder operates on features from the two pathways to produce detection results. Surprisingly, our proposed simple yet effective approach yields strong performance on three commonly used ISTD benchmarks. MI-DETR achieves 70.3% mAP@50 and 72.7% F1 on IRDST-H (+26.35 mAP@50 over the best multi-frame baseline), 98.0% mAP@50 on DAUB-R, and 88.3% mAP@50 on ITSDT-15K, demonstrating the effectiveness of biologically inspired motion-appearance integration. Code is available at https://github.com/nliu-25/MI-DETR.

MI-DETR: A Strong Baseline for Moving Infrared Small Target Detection with Bio-Inspired Motion Integration

TL;DR

A bio-inspired dual-pathway detector that processes one infrared frame per time step while explicitly modeling motion while explicitly modeling motion is proposed, demonstrating the effectiveness of biologically inspired motion-appearance integration.

Abstract

Infrared small target detection (ISTD) is challenging because tiny, low-contrast targets are easily obscured by complex and dynamic backgrounds. Conventional multi-frame approaches typically learn motion implicitly through deep neural networks, often requiring additional motion supervision or explicit alignment modules. We propose Motion Integration DETR (MI-DETR), a bio-inspired dual-pathway detector that processes one infrared frame per time step while explicitly modeling motion. First, a retina-inspired cellular automaton (RCA) converts raw frame sequences into a motion map defined on the same pixel grid as the appearance image, enabling parvocellular-like appearance and magnocellular-like motion pathways to be supervised by a single set of bounding boxes without extra motion labels or alignment operations. Second, a Parvocellular-Magnocellular Interconnection (PMI) Block facilitates bidirectional feature interaction between the two pathways, providing a biologically motivated intermediate interconnection mechanism. Finally, a RT-DETR decoder operates on features from the two pathways to produce detection results. Surprisingly, our proposed simple yet effective approach yields strong performance on three commonly used ISTD benchmarks. MI-DETR achieves 70.3% mAP@50 and 72.7% F1 on IRDST-H (+26.35 mAP@50 over the best multi-frame baseline), 98.0% mAP@50 on DAUB-R, and 88.3% mAP@50 on ITSDT-15K, demonstrating the effectiveness of biologically inspired motion-appearance integration. Code is available at https://github.com/nliu-25/MI-DETR.
Paper Structure (24 sections, 15 equations, 6 figures, 6 tables, 1 algorithm)

This paper contains 24 sections, 15 equations, 6 figures, 6 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Single-Frame Infrared Small Target Detection
  4. Multi-Frame and Motion-Based ISTD
  5. DETR-Based Object Detection
  6. The Constructive Nature of Visual Processing
  7. Methods
  8. Overview of MI-DETR
  9. Stage I: Retina-Inspired Cellular Automaton (RCA)
  10. Stage II: Parvocellular--Magnocellular Interconnection
  11. Stage III: High-Level Object Recognition
  12. Experiments
  13. Datasets and Evaluation Metrics
  14. Implementation Details
  15. Comparison with State-of-the-Art Methods
  16. ...and 9 more sections

Figures (6)

  • Figure 1: Performance comparison on the IRDST-H benchmark chen2025motion.
  • Figure 2: Parallel processing in visual pathways. The primate visual system separates motion and appearance signals in the retina, enables their interaction in V1 layer 4B, and integrates them in higher cortical areas for object recognition.
  • Figure 3: Overall architecture of MI-DETR.
  • Figure 4: Qualitative comparison on ITSDT-15K (confidence: 0.5, NMS: 0.3). Green boxes denote ground truth annotations, blue boxes indicate MI-DETR predictions, and red boxes show results from five multi-frame baselines (iMoPKL iMoPKL, SSTNet chen2024sstnet, STME peng2025moving, Tridos Duan_2024_Tridos, TMP zhu2024tmp).
  • Figure 5: Precision--Recall (PR) curves comparing 11 representative methods across three benchmarks: (a) ITSDT-15K, (b) DAUB-R, and (c) IRDST-H. MI-DETR consistently achieves superior precision-recall trade-offs across all datasets.
  • ...and 1 more figures