On the use of a scanpath predictor and convolutional neural network for blind image quality assessment

作者：

Highlights：

• A blind quality assessment metric for 2D images using deep learning.

• We proposed a saliency-based patch selection.

• Several pre-trained models were evaluated and compared.

• Evaluation of the saliency-based patch selection proposed to a random selection as well as the selection of all patches.

摘要

•A blind quality assessment metric for 2D images using deep learning.•We proposed a saliency-based patch selection.•Several pre-trained models were evaluated and compared.•Evaluation of the saliency-based patch selection proposed to a random selection as well as the selection of all patches.

论文关键词：Image quality,CNN model,Saliency,Scanpath prediction

论文评审过程：Received 30 October 2019, Revised 20 June 2020, Accepted 2 August 2020, Available online 15 August 2020, Version of Record 21 August 2020.

论文官网地址：https://doi.org/10.1016/j.image.2020.115963