QAC Index with pre-learned quality aware clusters for blind image quality assessment

Description

INPUT :

im:                         0~255 8bit grayscalae image;

cluster_center:       a cell array with structure for the quality aware cluster center;

cluster_center[l] is the cluster centers at the lth quality level, and it has the following domain:

centroids:          [Kxd double], K is the cluster number center at each quality level, d is the feature dimension.

quality:                     the quality for the lth level;

feature_fun:             the function handle for the patches based feature computation;

blksize:                    patch size, the default value is 8;

step:                        the step between patches, large step with speedup the  computation while maintain the same performance, the default step is 12;

OUTPUT :

Q:                              the predicted quality of the input image (im), range [0 1].

Blind Image Quality Prediction Using Joint Statistics of Gradient Magnitude and Laplacian Features

Efficient Marginal Likelihood Optimization in Blind Deconvolution