CVPR15 - OpenCV Vision Challenge.
Hi,
OpenCV is sponsoring a vision challenge in the upcoming CVPR convention:
http://code.opencv.org/projects/opencv/wiki/VisionChallenge
The challenge involves 11 benchmarks of various computer vision problems, with the goal to contribute state of the art algorithms (and code) to OpenCV.
Is anyone here thinks of participating? I'll be working on some of the "recognition" benchmarks.
Gil.
I would contribute some stuff, but mine is far from state of the art, rather functionality to make training and executing object models easier. But go try, who knows you might win!
My field is not part of the challenge. For the recognition challenges: state of the art is probably some form of deep neural network, or? Maybe then some deep learning stuff finds its way to OpenCV...
If it would be up to me, I would stay as far as possible from deep neural networks, they have indeed proven their succes, but the moment you want to know what it does internally, everyone pulls away their hands. All kind of neural networks are very difficult to debug once you want to find out what is going wrong, certainly if you have a lot of hidden neuron layers. This is the reason why keypoints / descriptors and learning techniques are still a valid alternative I think :)
@StevenPuttemans, thanks for the advice!