I will appreciate if I may find someone how can clarify for me the part regarding the quality of feature descriptor, shown in the figure below:
and this screenshot is from the article: SHREC
All my problem is that I'm not able to understand the situation clearly, appreciate simple words hoping to understand. Further indications, what I'm trying to do is mentioned in this question, but I wasn't able to get a fair answer for it. : Evaluation
So, it's about having same descriptor but for two models, main and scaled(or rotated, translated...) and trying to check how much my descriptor is good and well.
Any other recommendations regarding such situation are much appreciated,I'm implementing my work in python.