Image Blur Type Identification using Pattern Classification Approaches

This book introduces four different approaches for blur classification to categorize the blur in one of the three classes namely motion, out of focus and atmospheric turbulence blur. These methods utilize the appearance of blur patterns in frequency domain as patterns. Statistical, wavelet, ridgelet and curvelet features of blur patterns in frequency spectrum are extracted in these methods and a feed forward neural network is designed to obtain a mapping of the blurred images to blur category.