Speeding up LIP-Canny with CUDA programming