Aesthetics and comfortableness-based 3D image quality evaluation method

The invention relates to an aesthetics and comfortableness-based 3D image quality evaluation method. The method comprises the following steps of S1. extracting aesthetics characteristics and consistent aesthetics characteristics of left and right views of each 3D image in a training image set and a to-be-predicted image set, so as to acquire an aesthetics characteristic set F1; S2. extracting comfortableness characteristics of each 3D image in the training image set and the to-be-predicted image set, so as to acquire a comfortableness characteristic set F2; S3. taking all images in the training image set as a machine learning characteristic set T1 by combining the aesthetics characteristic set F1 and the comfortableness characteristic set F2, and training to obtain a 3D image quality evaluation model; and S4. evaluating each to-be-predicted image by using the trained quality evaluation model, so as to obtain final quality evaluation scores of all to-be-predicted images. The method canhelp make the evaluation