1：Experimental Design and Method Selection:

The method involves computing the nonlinear gradient magnitude with Gaussian smoothing of the reference and distorted images, constructing the relation intensity ratio and detail similarity between them, forming the final IQA map by linearly combining the relation intensity ratio with the detail similarity, and adopting a new pooling strategy to predict image quality.

2：Sample Selection and Data Sources:

Experiments are based on two publicly available databases (TID2008 and CSIQ).

3：List of Experimental Equipment and Materials:

Not explicitly mentioned.

4：Experimental Procedures and Operational Workflow:

The procedure includes computing the relation intensity ratio and detail similarity, forming the final IQA map, and using a new pooling strategy to predict image quality.

5：Data Analysis Methods:

The evaluation protocols are selected by Spearman rank correlation coefficients (SRCC) and PLCC.