Computer Vision: Real-time Pattern Recognition using C++

The algorithm recognizes patterns that are printed on a white sheet of paper. To detect the paper, it has to be placed on a darker background. It is possible to recognize over 100 different patterns. The steps of the algorithm include adaptive thresholding, detecting contours, histogram normalization and template matching. The algorithm is programmed in C++ and runs in real-time, using only a few milliseconds per frame. It is thus also applicable on embedded systems and cell phones.