Convolutional Neural Networks for No-Reference Image Quality Assessment

• Published in: 2014 IEEE Conference on Computer Vision and Pattern Recognition pp. 1733 - 1740
• Main Authors: Le Kang, Peng Ye, Yi Li, Doermann, David
• Format: Conference Proceeding Journal Article
• Language: English
• Published: IEEE 01.06.2014
• Subjects: Computer vision; Conferences; Convolutional Neural Network; Distortion; Feature extraction; Image quality; image quality assessment; Kernel; Networks; Neural networks; Nonlinear distortion; Optimization; Pattern recognition; Regression; Training; Transform coding
• ISSN: 1063-6919; 1063-6919
• DOI: 10.1109/CVPR.2014.224

In this work we describe a Convolutional Neural Network (CNN) to accurately predict image quality without a reference image. Taking image patches as input, the CNN works in the spatial domain without using hand-crafted features that are employed by most previous methods. The network consists of one convolutional layer with max and min pooling, two fully connected layers and an output node. Within the network structure, feature learning and regression are integrated into one optimization process, which leads to a more effective model for estimating image quality. This approach achieves state of the art performance on the LIVE dataset and shows excellent generalization ability in cross dataset experiments. Further experiments on images with local distortions demonstrate the local quality estimation ability of our CNN, which is rarely reported in previous literature.