On their site it is marked that OpenBR is using OpenCV, so your approach may be not so good because of your database, or because of the implementation optimisation; or maybe it could be faster because you are a ninja.
i don't think, their lbp /ltp implementations are faster(rather not) or better(depends on classifier used) than opencv's but again, they got some things, opencv does not have , like e.g. this one
we'll never know, unless someone goes there and does actual tests. ;)
So, again their stuff may be better, because there is a team (I suppose) that worked on that thing, and optimise it, and doing the application your own (using OpenCV) it will be almost certain that it will not be as good as theirs.
@thdrksdfthmn thank you very much for your concern. But I just wonder if OpenBR does not share most of its algorithms inherent to face stuff with OpenCV ?
IMHO, it uses face detection, then features/descriptors and compare them somehow with something, that I do not thing is implemented in OpenCV algorithms. So, they must have a database of their owns based on which they have computed some sort of algorithm that is telling the age, etc; but they used the algos in OpenCV for getting the info they need to process. You can get deeper in their code to find out exactly
On their site it is marked that OpenBR is using OpenCV, so your approach may be not so good because of your database, or because of the implementation optimisation; or maybe it could be faster because you are a ninja.
if you could do some tests, and come back here with results, much appreciated ;)
@thdrksdfthmn do you mean that OpenBR is faster and more efficient than OpenCV in different facial detection ?
Detection, I do not think so, but maybe in feature interpretation will be faster than your implementation...
i don't think, their lbp /ltp implementations are faster(rather not) or better(depends on classifier used) than opencv's but again, they got some things, opencv does not have , like e.g. this one
we'll never know, unless someone goes there and does actual tests. ;)
So, again their stuff may be better, because there is a team (I suppose) that worked on that thing, and optimise it, and doing the application your own (using OpenCV) it will be almost certain that it will not be as good as theirs.
@thdrksdfthmn thank you very much for your concern. But I just wonder if OpenBR does not share most of its algorithms inherent to face stuff with OpenCV ?
IMHO, it uses face detection, then features/descriptors and compare them somehow with something, that I do not thing is implemented in OpenCV algorithms. So, they must have a database of their owns based on which they have computed some sort of algorithm that is telling the age, etc; but they used the algos in OpenCV for getting the info they need to process. You can get deeper in their code to find out exactly