Channel Recombination and Projection Network for Blind Image Quality Measurement

In this article, a novel channel recombination and projection network (CRPNet) is proposed for no-reference image quality assessment (NR-IQA). The proposed CRPNet is composed of four parts, a feature extractor, a channel recombination strategy (CRS), a saliency-guided selective projection (SSP), and a channel score weighting (CSW). The feature extractor is first utilized to learn patch-level features from patches to address the lack of training samples. The CRS is proposed to obtain image-level features by recombining patch-level features in channel dimension, which can solve the mismatch problem between the image score and the patch-level quality. To further increase the image quality prediction accuracy, the SSP is designed, in which the mapping ratios of fully connected layers are calculated by the saliency priority of patches. Moreover, a CSW is introduced to enhance the image visual quality representation. Experimental results on six IQA datasets, e.g., Laboratory for Image and Video Engineering database (LIVE), Tampere Image Database 2013 (TID2013), Categorical Subjective Image Quality database (CSIQ), LIVE Multiply Distorted image quality database (LIVE MD), LIVE in the wild image quality CHallenge database (LIVE CH), and Waterloo Exploration Database, show that the proposed CRPNet outperforms the state-of-the-art NR-IQA methods, and achieves the superior performance. Code is available at https://github.com/zhaob10/CRPNet.