Image and Video Sensor Modality Extension of the Real-Time Streaming H-Pipeline: A 140-Study Validation Campaign Anchoring Per-Frame Feature Embeddings via Convolutional Neural Networks and Vision Transformers for Privacy-Preserving Smart-Glasses, Bo

v2 (2026-06): Corrected attribution-- compressibility and privacy are now attributed to the codebook-quantization layer, lossless invertibility to the spiral front-end. All numerical results unchanged. This deposit accompanies the v1 manuscript for Paper 15 of the spiral-domain encoder validation campaign series. It contains the complete pre-registration scope document, runner code, summary verdicts, POC addenda, and arXiv-format manuscript for a 140-study empirical validation campaign of the image and video sensor modality extension of the real-time streaming H-pipeline architecture introduced in Paper 14. The campaign comprises 140 studies executed in a continuous deterministic protocol (PYTHONHASHSEED=0) across eighteen thematic sections: §1–§9 pre-registered (50 studies at 88.0% SUPPORTED) and §10–§18 post-hoc-disclosed (90 studies at 92.2% SUPPORTED). The full campaign achieves 90.7% SUPPORTED (127 of 140). The v1 manuscript is text-only; figures are planned for a v2 manuscript. Headline findings · Canonical image/video operating point validated: window W=1.0s, hop H=0.5s, frame rate fs=30Hz, spiral d-axis=5, D-input=768. The d-axis=5 / D-input range covers ViT-Tiny D=192 through ResNet-50 D=2048 per-frame feature embeddings. · Per-frame feature embedding pipeline: CNN/ViT extractor → spiral encoder → codebook quantization → streaming H-pipeline (Stages L-A through L-D). · Capable face-recognition MIA AUC envelope 0.55-0.65 preserved via Stage K-1 identity-axis projection (Parent K composition). · AR(1) inter-frame correlation in synthetic image-feature streams replicates real-world temporal smoothing observed in deployment. · FDA SaMD Class I, II, and III compatibility across 4 vision embodiments: M smart-glasses, N body-worn camera, O continuous medical imaging, P surveillance. · §17 raw-pixel 2D spatial mode boundary disclosure (Studies 1466-1475): the inventors characterize raw-pixel mode as duplicating Vision Transformer (Dosovitskiy et al. 2020) and 1980s raster-scan patch-encoding prior art without advancing energy / compression / privacy goals beyond Parent L v5 per-frame feature-embedding mode. Honest boundary disclosure documenting why raw-pixel mode is NOT pursued as a separate Parent. Related provisional patent application Parent L (U.S. Provisional Patent Application 64/084,827, filed at USPTO on 2026-06-11; 70 claims, 56 pages). The image/video modality content is encoded in Parent L through a canonical-parameter table extension, 4 new vision embodiments (M smart-glasses, N body-worn camera, O continuous medical imaging, P surveillance), 4 new dependent claims (65-68), and Limitations 14-15. Three companion standalone provisionals filed the same day: Parent I (64/084,807) federated, Parent J (64/084,817) codebook-centroid privacy hardening, Parent K (64/084,821) pre-encoder identity-axis projection. Companion deposits in the campaign series · Paper 8 — Phase XXI musculoskeletal-DT validation: 10.5281/zenodo.20466035 · Paper 9 — Phase XXII surgical-robot patient-pair DT: 10.5281/zenodo.20470763 · Paper 10 — Phase XXIII unidirectional release-only bioelectric DT: 10.5281/zenodo.20480183 · Paper 11 — Phase XXIV bidirectional closed-loop bioelectric DT: 10.5281/zenodo.20495976 · Paper 12 — Phase XXV IoT-LLM pipeline: 10.5281/zenodo.20499453 · Paper 13 — Phase XXVI federated multi-institutional (this deposit cycle) · Paper 14 — Phase XXVII streaming + image/video extension (this deposit cycle) Reproducibility All 140 studies execute deterministically under PYTHONHASHSEED=0. The reproducibility corpus includes the vision-modality streaming harness library and 140 per-study summary JSONs. End-to-end campaign reproduction wall-clock: ~8 minutes on Apple M2 / 16 GB. License Creative Commons Attribution 4.0 International (CC-BY 4.0). Commercial deployment of the image/video streaming H-pipeline requires a license to Fieldstone Analytics, LLC. Contact: [email protected]. Citation Ferlic, R.J. and Ferlic, K.K. (2026). Image and Video Sensor Modality Extension of the Real-Time Streaming H-Pipeline: A 140-Study Validation Campaign. Zenodo. https://doi.org/10.5281/zenodo.20596931