Ok, so first we had automatic exposure, then focus, then face detection and now smile detection. Why not step up a bit and do full AUTO? The camera reads the frame buffer, applies some computations and if certain parameters are met, takes a picture
I'm thinking the camera should be level (horizontal or vertical), in focus, image should not be blurred, horizon/split line must be at one of the thirds and various other composition rules: minimum contrast, maximum range of colors so the picture is not too busy, etc.
We have access to the frame buffer (motion detection), can do some processing (edge detection, contrast detection, metering, etc) so all the tools are there.
Various thresholds could be set to enforce some rules more or less.
Signal to noise ratio will be the same, there are many shutter-happy photographers out there
This is essentially motion detection with different processing. Even at 0.1fps processing speed it should still yield some interesting results while you walk with the camera hanging from your neck
