RailSyn: Diagnosis-Guided Image Generation for Traceable Data Completion in Railway Foreign Object Detection
2026-08-31 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors focus on improving the detection of foreign objects on railways, which is important for safety but difficult because real examples of such objects are rare and varied. They created RailSyn, a two-part system where an Inspector analyzes what real data is missing and a Generator creates synthetic images to fill those gaps. This approach helps make the detection models better at recognizing different kinds of objects and conditions. Tests show that using RailSyn improves detection accuracy across many popular methods.
foreign object detectionrailway safetysynthetic data augmentationdomain adaptationempirical coverrepresentation spaceobject placementconditional refinementaverage precision (AP)
Authors
Quan Hao, Chenxi Zhang, Ziyang Tao, Yuyuan Zhou, Yudong Wang, Rui Shi, Lechuan Xu, Changhao Liu, Liguo Zhang
Abstract
Railway foreign object detection (RFOD) is critical to safe railway operation, yet scarce real positive samples incompletely represent task-relevant variations in object scale, intrusion relation, railway scene, illumination, and adverse weather. Existing synthetic augmentation can improve RFOD detection, but its gains lack an explicit account of the task-relevant deficiencies complemented by the generated data. We therefore introduce RailSyn, a diagnosis-guided framework comprising a real-referenced Inspector and a requirement-aligned Generator. The Inspector constructs a variable-radius empirical cover from finite real observations to localize candidate completion regions and profile synthetic pools. The resulting audit identifies railway-context, intrusion-semantic, and visual-consistency requirements; the Generator addresses them through domain adaptation, agent-planned placement and physical contact relations, and plan-consistent conditional refinement. Using the Inspector, we further trace representation-space changes across generation variants; the complete system attains a local-shell occupation of $C_{gap}$ to 13.64%, which measures generated coverage of real-derived completion regions. Extensive experiments show AP50--95 gains of up to 4.9 points and consistent improvements across nine mainstream detectors, demonstrating broad cross-architecture utility.