Transient Optical Detection Transition (TODT):A Pilot State-Space Analysis of a Falling Quartz Object in a Natural Indoor Scene
We present a pilot framework for analyzing transient loss of visual detectability of a physically present, moving object without invoking physical transparency. A ~2 cm quartz stone was dropped from approximately 2 m against a natural indoor background and recorded in a smartphone slow-motion workflow. A short sequence in the trimmed video contained directly visible compact image evidence consistent with the falling quartz; subsequent positions were constrained by a ballistic continuation model and local image evidence. The measured local scene contrast along the reconstructed path was then propagated through a simplified human-vision model incorporating angular size, residual retinal slip, spatial contrast sensitivity, optical/retinal blur, temporal integration, and pursuit limits. Under a strict Monte Carlo observer family (20,000 samples; detection threshold 0.5-1.0%, pursuit capacity 60-100 deg/s, integration time 20-40 ms), the inferred state changed from VISIBLE to SUBTHRESHOLD at frame 443 and then to TRANSITION at frame 444. At frame 443, local scene contrast was 1.86%, median effective contrast was 0.256%, and the modeled probability of falling below the detection threshold was 97.7%. The longest marginal/subthreshold run was 33.3 ms. These results do not demonstrate physical transparency or prove subjective invisibility. Instead, they support a testable state-space hypothesis: a real moving object can transiently enter a model-defined subthreshold visual state when local background contrast collapse and motion-dependent retinal processing coincide. Controlled psychophysics and independently calibrated eye tracking are required for validation. Keywords: visual detectability; contrast sensitivity; retinal image motion; state transition; motion perception; natural scenes
Authors
- Osuke Doijiri
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-08-21
- DOI
- https://doi.org/10.5281/zenodo.22043584
- Primary Topic
- Visual Attention and Saliency Detection
- Type
- preprint