This is the repository for my android app AutoFocus
This app uses computer vision algorithm (or more specifically image signature now) to predict the region which draws most attention from the subject, or salient region.
Currently, we are only uses image signature algorithm, which combines three channel (in opponent color space). In the future, I will add more visual cues to predict the salient region more accurately, e.g., face detection, pedestrain detection.
For question, please contact me at zhangtemplar@gmail.com