Comment by nickmcc

I wonder if you develop an ML model that uses brightness and color reduction cues in the video in order to estimate depth in lieu of a dive computer. I suspect cameras auto-exposure or underwater modes may nullify this.