Image Blur Classification and Estimate Parameter via Deep Learning

Image blur kernel classification and parameter estimation are critical for blind image deblurring. Current dominant approaches use handcrafted blur features that are optimized for a certain type of blur, which is not applicable in real blind deconvolution application when the Point Spread Function (PSF) of the blur is unknown. In this paper, a Twostage system using Deep Neural Network (DNN) and General Regression Neural Network. The proposed approach is to first classify the blur type using DNN and then identify its parameters using GRNN and DNN has been applied to the problem of blur analysis. In the blur type classification, this method attempts to identify the blur type from mixed input of various blurs with different parameters, rather than blur estimation based on the assumption of a single blur type in current methodology and then classify those features. Moreover, in the parameter identification, the edge detection on logarithm spectrum helps DNN to identify the blur parameters with very high accuracy. Experiments demonstrate the effectiveness of the proposed methods with better results compared to the abundant dataset. Keywords— Blur Kernel classification, Point Spread Function, Deep Neural Network, General Regression Neural Network.