If I have now identified a video face region (area) can also be delineated specific coordinates, I wonder when reading video files to first identify which is the human face area (delineating specific location), and to deal with this streaming video data, opencv can do?

I actually is the easiest to do identify faces) part of (or a specific location, and then the poor quality of network transmission only discard the rest of the data, make video playback time background flowers screen but does not affect facial spot (area) to watch the effect, don't know openCV can do it, or any other tool can do?

I see most of them are directly extracted from a particular area and form independent files, or in a particular area draw mark, I have to do is at the time of reading processed video files, which can identify which is the civilian regional data flow, and then directly processing, don't know bosses have what way?

CodePudding user response:

ROI

CodePudding user response:

reference 1/f, zhao teacher reply:

ROI

I understand the function of the ROI, now I need to know is, when I read the file in flow order flow (or data flow), how do I know that I read the packet belongs to ROI area, not only is the use of certain tools ROI marked out for the human eye to distinguish, I now need to packet level judgment!

CodePudding user response:

5 x 4 pictures and ROI is as follows:
AAAAA
B ROI
BBC ROI CC
DDDDD
Corresponding flow:
AAAAAB BBC ROI ROI CCDDDDD
.

CodePudding user response:

quoted zhao teacher reply: 3/f

5 x 4 pictures and ROI is as follows:
AAAAA
B ROI
BBC ROI CC
DDDDD
Corresponding flow:
AAAAAB BBC ROI ROI CCDDDDD
.

your meaning similar to do a matching algorithm of sliding window?

CodePudding user response:

Through pass by look at the past