When the author talked about scanning the video for "interesting" parts he mentioned pattern recognition. Does anyone know if data compression is used for this in applications? The idea being that if you have 12 hours of cctv footage that is mostly just the same view of a warehouse most of that will compress very well but not the 15 or 20 minutes where there were people walking around.
Never mind hooking up to your fast PC, just add 4G+ network access and have it stream the video straight to the cloud for immediate analysis and indexing.
I just thought it was cool that he was basically describing what HighlightCam is doing. Nice validation from an entrepreneur who has had a big success.
