Hi all,
I have to track an object using LKoptical flow for that I require a set of good features in the specified region. Here in the forum I got example how to use LK and I also used the same in C++. But my question is, in C++ opencv I used a function called setROI for the image to find out the good features in the particular region. But If I use corner detection instead of good feature detection here in FastCV, I have to use it for whole image instead of specific region in the given image. Please will anyone tell me is that functionality available in fastcv? If it is will u tell me which api it is?. If thats not possible can anyone suggest me to use anyof the other ways to get the things done.
Thanks in advance,
Cheers,
Shravan
Hi Shravan,
In upcoming June release of FastCV, ROI will be supported through use of the stride parameter in applicable function APIs. For your information, stride is the number of bytes for one row of image (or 2D array) including any necessary padding for memory alignment. So ROI in FastCV is accomplished differently than OpenCV (i.e. no SetImageROI call) but it's function equivalent.
Cheers,
-Jeff