Full-Reference Image Quality Assessment with Transformer and DISTS

To improve data transmission efficiency, image compression is a commonly used method with the disadvantage of accompanying distortion. There are many restoration (IR) algorithms, and one most advanced algorithms generative adversarial network (GAN)-based high correlation to human visual system (HVS). evaluate performance GAN-based IR we proposed an ensemble quality assessment (IQA) called ATDIQA (Auxiliary Transformer DISTS IQA) give weights on multiscale features global self-attention transformers local convolutional neural (CNN) IQA DISTS. The result not only performed better perceptual processing (PIPAL) dataset images by GAN but also has good model generalization over LIVE TID2013 as traditional distorted datasets. successfully demonstrates its judgment score images.