An optimized CNN-based quality assessment model for screen content image

作者：

Highlights：

• A valid network architecture is designed for both NR and FR SCI quality evaluation.

• A training data selection method is proposed to fine-tune the pre-trained model.

• An adaptive pooling approach is employed to fuse patch quality, owns strong noise robust.

摘要

•A valid network architecture is designed for both NR and FR SCI quality evaluation.•A training data selection method is proposed to fine-tune the pre-trained model.•An adaptive pooling approach is employed to fuse patch quality, owns strong noise robust.

论文关键词：Image quality assessment,Screen content image,No-reference,Full-reference,Convolutional neural network,Quality pooling,Data selection

论文评审过程：Received 24 March 2019, Revised 8 August 2020, Accepted 24 January 2021, Available online 6 February 2021, Version of Record 11 February 2021.

论文官网地址：https://doi.org/10.1016/j.image.2021.116181