Hallucinated-IQA: No-Reference Image Quality Assessment via Adversarial Learning

直接引用P1-2:“The ill-posed nature of the underdetermined NR-IQA problem is particularly pronounced for the limited information,for which the form of distortion and the corresponding non-distorted reference image are typically absent.”P2:“a high-resolution scene hallucination is firstly generated from the distorted image. Then, the discrepancy map which naturally encoding the difference between the distorted images and hallucinated reference can be obtained to guide the learning of regression network”P5:“Even when G fails to generate high-resolution hallucination images, the predicted scores of R should still be