Creating individual images from detected ROI objects

[zinon]

Hi,

I'm pretty new using the IPP library and Ido not know how, from an image with multiple objects, detect each one and create a individual image for each detected object. My image is B&W with white background and grayscale obejcts.

I know thatit is possible to get an ROI pointer and set its width and height, and with that I can copy the ROI. But how can I get all the objects collection from the original image and get each objects (or ROI) pointer and then copy or cut it?

Thanks!

[Ying_H_Intel]

Hi zinon,

Your problem looks more image detect or image segementationthanIPP usage problem.You may consider implement some image detect algorithms(which may call some ipp functions like ippiBoundSegments() in ipp computer vision domain). Once you get the ROI's position, you can use ippicopy() copyany ROI from original image.

IPPprovide one sample named UIC Sample Demo (andsource code in image-codecsuicsrc)
) <http://software.intel.com/en-us/articles/intel-integrated-performance-primitives-intel-ipp-intel-ipp-sample-code/>.
It can do face detect and simply image segemention. You may take a look at it first and get the idea how to implement a algorithm based on ipp functions.

Best Regards,
Ying