Motion-Induced Visual Primitives from Commodity Smartphone Video

VeilSight Motion Glyph Dataset — v3.0 Cross-Video Persistence, Scale Robustness, and Domain-Specific Motion Structure Overview This release (v3.0) extends the VeilSight Motion Glyph Dataset with new analyses demonstrating cross-video persistence, scale robustness, and domain specificity of motion-induced visual glyphs extracted from unconstrained smartphone video. Motion glyphs are treated as emergent, phase-structured motion primitives rather than object-centric features. They arise from consistent spatiotemporal dynamics in natural video and persist independently of scene content, camera parameters, or recording instance. Version 3.0 focuses on validating glyph identity across independent recordings and testing robustness under temporal rescaling, providing evidence that these primitives function as stable dynamical attractors rather than dataset artifacts. What’s New in v3.0 1. Cross-Video Glyph Persistence Glyphs are evaluated across multiple independent videos capturing the same natural motion domain. For each glyph, we report: Number of distinct videos in which the glyph appears Total frames observed across videos Mean and variance of stability metrics Mean entropy across recordings Result:Each identified glyph recurs consistently across all sampled videos, with low inter-video variance in stability and entropy. This demonstrates that glyphs are not video-specific and do not arise from incidental recording noise. 2. Scale Robustness Analysis Glyph dynamics are evaluated under temporal subsampling to assess sensitivity to frame rate and sequence length. Metrics include: Change in stability under scale reduction Change in entropy under scale reduction Baseline-normalized deltas Result:Glyph identities remain stable under temporal scaling, indicating that they capture structural motion properties rather than fragile timing coincidences. 3. Domain-Specific Dynamics (Tree vs Non-Tree Comparison) A controlled comparison is performed between: Videos from the target natural motion domain Non-domain baseline video Metrics include: Mean and median glyph stability Self-dynamics dominance Total transition activity Result:Structured glyph dynamics emerge only within the target domain. Non-domain video collapses to trivial or null dynamics, confirming that glyph structure is content-conditional, not algorithmically imposed. Dataset Contents This release includes: Aggregated glyph-level statistics Cross-video persistence metrics Scale robustness metrics Domain comparison summaries All data are provided as CSV files suitable for statistical analysis and downstream modeling. No raw video is included. Intended Use This dataset is intended for: Research in motion representation and temporal structure Studies of emergent dynamics in natural video Comparative evaluation of motion primitives Exploratory work on symbolic or pre-symbolic visual representations The dataset is not intended to function as a drop-in replacement for object detection, tracking, or supervised action recognition benchmarks. Reproducibility Notes The analyses provided in this release are derived from deterministic processing pipelines. While full internal system details are not disclosed, the published outputs are sufficient to support: Independent statistical verification Comparative benchmarking Theoretical analysis of motion stability and entropy Future releases may expand the number of videos and domains evaluated. Version History v1.0 — Initial release of motion glyph primitives v2.0 — Phase-structured glyph aggregation and omega-distance analysis v3.0 — Cross-video persistence, scale robustness, and domain specificity Citation If you use this dataset or analyses in your work, please cite: Rice, D. (2025). VeilSight Motion Glyph Dataset (v3.0). Zenodo.https://doi.org/10.5281/zenodo.XXXXXXX (Replace with the DOI assigned to this version.) Notes on Scope and Disclosure This release intentionally focuses on empirical properties of motion glyphs rather than internal extraction mechanisms. The goal is to enable scientific evaluation of glyph behavior while preserving system integrity.