I am searching for Feature Descriptors an as I found a keypoint detector for Intrinsic Shape Signature (ISS), but I can not find any implementation of the corresponding descriptor. This is a little confusing to me, because in the keypoint detectors source code is a reference to the original ISS paper, which mostly refers to the descriptor and descriptor database indexing for faster matching.
I am wondering if there is an implementation of the ISS descriptor too and if not, is there reason for it, except that no one has done it since yet?
thanks for your replay, but I am searching for a way to obtain the ISS descriptors. I know the class and usage of the keypoint detector since I read the docs and paper to the approach. What I am looking for is something like a ISSEstimator or something like that, to compute the ISS histograms and reference frames like described in , also referenced in the keypoint detector class ISSKeypoint3D.
The descriptor an keypoints of the Yu Zhong's approach seem to be nice for recognition tasks, but at the moment I am not able to calculate the ISS descriptor itself. Therefore I want to know if the pcl contains an implementation of it that I cannot find, or if I have to implement it myself if I want ISS.
 Yu Zhong, “Intrinsic shape signatures: A shape descriptor for 3D object recognition,”
Computer Vision Workshops (ICCV Workshops), 2009 IEEE 12th International Conference on ,
vol., no., pp.689-696, Sept. 27 2009-Oct. 4 2009
Ok, thanks for your effort. However, as I already mentioned: I know the keypoint detector, but I do not know how to acquire the histograms encoding the surface information.
Nevertheless was this a good hint, because I didn't expect to get the scatter matrix to estimate the local reference frame from the keypoint detector. Thank you for that one.
I still cannot compare any local feature descriptors, since it does not seem like there are no histograms available in the keypoint detector .