Computer vision lab
Six hands-on modules covering processing, detection, inspection, AR, face analysis and visual search, with the business case for each.
Runs 100% in your browser. No sign-up, no backend. Models download once and are cached; your images never leave your device.
Build classical pipelines, inspect histograms, and segment with k-means or SegFormer.
DETR detection, top-5 classification, CLIP zero-shot labels, and live webcam counting.
Golden-reference differencing, blob finding, anomaly scoring and SPC dashboards.
Face mesh try-on, hand gestures that drive UI, and approximate measurement.
Consent-first detection, liveness hints, 1:1 verification and FAR/FRR trade-offs.
Shop-the-look with CLIP embeddings, text + image refinement, and an embedding map.
Capture
Upload, paste or use your webcam.
Process
Classical kernels clean and transform pixels.
Infer
Models run on WebGPU or WebAssembly locally.
Explain
Concept cards show how and why it works.