context-aware encoding

[liujianisme]

Hi, thanks for the release. I'm currently working on context-aware video encoding which aims at setting different QPs for different area within frame to be encoded according to grayscale saliency map where white areas indicate ROI areas. Now I'm struggling at how to implement context-aware encoding in V4L2. Any advices would be very appritiated.

[azeam]

Sounds like an interesting project, but the short answer would be "no idea", I've never looked into those areas.

[liujianisme]

Hi, thanks for reply. I know there are extradata roi input in video encoding settings, could it be possible to implement saliency map using extradata roi input?

[suuuehgi]

This issue has nothing to do with the camset.