Reproducibility package for Frame Blur as an Image-Plane Motion Cue in Counter-UAV and Drone-Tracking Video

This dataset provides the reproducibility package for the manuscript “Frame Blur as an Image-Plane Motion Cue in Counter-UAV and Drone-Tracking Video”. The study tests whether single-frame target blur provides a reliable image-plane motion cue in real counter-UAV and drone-tracking video. Across the studied footage, measured frame-blur length does not show a camera-consistent operational association with motion, while controlled-exposure experiments confirm estimator response only above the observed real-footage range. The package contains Python source code, processed numerical results, diagnostic outputs, environment records, execution commands, file manifests, and SHA-256 checksums supporting the reported analyses, tables, and supplementary results. It covers the primary and alternative blur estimators, bootstrap and robustness analyses, controlled-exposure experiments, temporal-averaging probes, and controlled prediction-sensitivity diagnostics. The original benchmark videos are not redistributed. The analyses use the publicly available Anti-UAV-RGBT, Multi-View Drone Tracking (MVDT), and FRED datasets, which must be obtained from their original sources. Users should begin with reproducibility/REPRODUCIBILITY.md and verify the included SHA-256 manifests before running the scripts.