Paper Cut-Out Patterns Recognition Based on Wavelet Decomposition and NMI

This paper proposes a new Paper Cut-Out Patterns Recognition method combining the Wavelet Transform with NMI feature. The NMI feature of binary image has the ability of anti-gray distortion and TRS invariance, and its extraction method is sample. By Wavelet Transform, we can compute the NMI feature using low-frequency components instead of the original image. Experiments show that it can remove noise effectively and a relatively good recognition performance could be achieved to recognize paper cut-out patterns of exaggerative deformations.