Blur Parameters Estimation using SUMFC and Wavelet Transform

A new blur parameters estimation algorithm is proposed in this paper. This scheme uses SUMFC for featureextraction. In order to the best performance of the algorithm, wavelet transform is used, which can determine whetheran image is blurred or increased and to what extent an image is blurred. To predict the blur length of a particularblurred image, we employ an RBFNN with combination of SUMFC of the corresponding blurred image and blur extentparameter which comes from wavelet transform as its input. With this approach we can estimate blur parameters inblurred images more accurately. The results reveal the ability of our method in predicting blur parameters

Keywords:

Motion Blur , Blur Extent , Wavelet Transform , Sum of the Magnitude of Fourier Coefficient , RadialBasis Function Neural Network