OpenCV detecting an application frame from a set of screenshots -
I have this problem, I want to find a special window with a set of screenshots. If I have 100 screenshots then I have to know what is there in iTunes. When I come to OpenCV, then am I still a Noah, do I have to train a Higher file for this Or is there to make or is there no easy solution because the app frames are not different?
The specific route you take will depend heavily on the details of your problem:
- Are you always looking for iTunes?
- How is it right to be?
- Is false positives acceptable? How about negative negative?
- Is there always a screenshot from the same system?
- Will iTunes be always maximized?
- Will desktop background ever change?
- iTunes should be identified in different circumstances, eg coverflow, visualization, iTunes store?
- Will iTunes always look the same?
For starters, I can see that you can see full simplified metrics and see if they separate your iTunes from non-iTunes images. For example, iTunes images have more white pixels than any other, in that case you can only count gray pixels
If that does not work, consider creating histograms of colors . Probably iTunes is used for most of the same color (gray, white, light blue and green). Then you can create different color histogram libraries in accordance with iTunes Use Cases (Playlists, iTunes Store, Coverflow, etc.) and can identify pictures of iTunes by matching histogram. See this example from this example
Comments
Post a Comment