Ask HN: Using AI to identify natural flow of a video interview
I am a somewhat of a noob when it comes to video editing — but I am masterful at software development, and would like find (or create) a technology that allows one to figure out during a video interview which scenes were edited to placed later in a video.
For example, during an interview — like the one that vice president Harris had with CBS recently — it should be possible to use the gloss degradation on the lips, or anything that is directional with time to trace back and identify how much time passed between scenes. Below are examples of things that are directional with time:
1. Lip gloss wearing out 2. Anything dependent on gravity 3. Individual hair strands becoming more chaotic 4. If shot in a place with a bit of natural light, the slight contrast shifts minute by minute 5. The level of water on a cup 6. The strain on the eye 7. Folds in the clothes that add up over time
I am curious if therea are any tools (or if any tools can be built) to help with this issue, not only for the interview, but at the grander scheme of things — this is a big issue in video editing and video analysis.
Any points or guidance in the right dirction would be appreciated.