implementing ORB in opencv c++
hi everyone;
for my final year project, i need to implement ORB ,ther is no tutorial available in the internet. I need your help, ideas , how to proceed and so on.
thank you in advance
maybe .
please clarify, if you're to reimplement ORB, or if you're just trying to use opencv's detectors/descriptors.
hi again,
yeah, i have to reimplement orb in c++ and opencv . now i am trying to understand the original code of orb in opencv which is quite difficult, you know, to understand a code of some one, i think that you have first to understand how this person thinks, lol. any way, i would appreciate any idea or help concerning this.
thanks again.
why reimplement something that already exists in the library using that library?
bacause simply my project consists of parallelizing this algorithme, and to master this task, it would be better to implement your own code, i hope that you have gotten my idea.
and what makes you believe that the actual code isn't parallelized and highly optimized, like all openCV code?
the code is in front of me, it is implemented in c++. if you want tp parallelize, then it depends in your machine ( and of course in your code), i mean the architecture, how many processors, how many cores, is there a shared memory , can we use SIMD .....and a lot of things. the original one is just the implementation of the algorithme , of course some optimizations have been taken into consideration to make the algorithme faster than sift and surf but IT HAS NOT BEEN PARALLELIZED.
afaik the feature extractors make use of the parallized inner for loops... which are highly optimized for multicore processing.
could you please explain me more and give me an exemple.
no i can't ... i am not skilled in these optimizations ... you will need a programmer of the algorithms for that!