I have a 2D Array, where each row is a feature vector.
Each feature vector is like: [x1 y1 x2 y2 x3 y3 x4 y4] where, (x1,y1) are coordinates of first feature point, (x2, y2) of second feature point and so on...
I need to give the feature points as input to a neural network. Now, how do I combine x1 and y1 into one meaningful unit before feeding it into neural network? I am stuck here...could anyone please suggest what to do?
I read somewhere we have to use some two-dimensional transformation...like SVD...but I can#t figure out how actually???
I am working with C++ and OpenCV 2.3
Any help would be very much appreciated...thanks!!!