Fig. 2

The implementation of blind blur assessment in MRI images. a The test image has its pixel intensity level rescaled to lie between 0 and 255. b Foreground of the test image in a is extracted. c The identical edge map from the initial parameters of the low and high energy difference of Gaussian filters. d The output image of the low energy filter at the conclusion of the multiscale representation. e The output image of the high energy filter at the conclusion of the multiscale representation. f The edge map extracted from the image in d. g The edge map extracted from the image in e. h Variation of image features from the output of the low and high energy filters at different Gaussian scales. i The predicted contrast, sharpness and total blur quality scores based on the analysis of the plot in h