image processing - what is the meaning and use of class member class_id of class cv::KeyPoint in opencv? -


in reference manual of opencv 2.4.3 keypoint::class_id described "object id can used clustered keypoints object belong to".

due lack of adequate knowledge 'keypoint' can't understand whats purpose of class_id.

my question is, using feature detector, descriptor , matcher can detect matching keypoints of query object in training image. how can segment out query object in training image. can use grabcut or watershed algorithm? if yes, how?

answer of either question helpful.

thanks in advance...

the class_id member field can used, in practice, store information find useful each keypoint. documentation says, can store, example, id of detected object.

for example, have image, extract keypoints (e.g. surf), run object detector these features, , result each input feature contains in class_id id of corresponding object, or -1 if attached no object. don't know if field filled opencv function or use.

regarding second question, matching features may not enough segment out object, because features can located inside object, don't information outline. starting point, though, if have more information object. example, if know object planar (a poster on wall), can use feature matches compute homography between input , training images. give outline. or, example, large collection of close matched features may indication of presence of object. that, may try other color segmentation, edge matching, etc.


Comments

Popular posts from this blog

basic authentication with http post params android -

vb.net - Virtual Keyboard commands -

How to get multiresult with multicondition in Sql Server -