How can one use artificial neural networks for video processing ?
Thank you for your question!
There are several ways to do this. The simplest one is using open source computer vision libraries like OpenCV. You just process video frame by frame. That is what we had used at the very beginning to recognize car plates. But it isn't CNNs.
The next step was using and training artificial neural networks like SegNet (
http://mi.eng.cam.ac.uk/projects/segnet/). It lets to recognize different objects in the video. The main point here is training of NN. You need a huge amount of content for each object you train to recognize. One NN trains for one object. We can provide you with some examples of our code showing how we use it.
Finally we are moving to TenzorFlow (
https://www.tensorflow.org/). I think we will write a big article about the TensorFlow using soon.
