Image Quality Assessment by Comparing CNN Features between Images
Authors:
- 1 November 2016
- journal article
- research article
- Published by Society for Imaging Science & Technology in Journal of Imaging Science and Technology
- Vol. 60 (6) , 060410-1
Abstract
Finding an objective image quality metric that matches the subjective quality has always been a challenging task. We propose a new full reference image quality metric based on features extracted from Convolutional Neural Networks (CNNs). Using a pre-trained AlexNet model, we extract feature maps of the test and reference images at multiple layers, and compare their feature similarity at each layer. Such similarity scores are then pooled across layers to obtain an overall quality value. Experimental results on four state-of-the-art databases show that our metric is either on par or outperforms 10 other state-of-the-art metrics, demonstrating that CNN features at multiple levels are superior to handcrafted features used in most image quality metrics in capturing aspects that matter for discriminative perception.Keywords
Funding Information
- Deutscher Akademischer Austauschdienst
This publication has 28 references indexed in Scilit:
- Objective assessment of perceived sharpness of projection displays with a calibrated camera2015
- Image database TID2013: Peculiarities, results and perspectivesSignal Processing: Image Communication, 2015
- Spatial pooling for measuring color printing quality attributesJournal of Visual Communication and Image Representation, 2012
- PHOG-Derived Aesthetic Measures Applied to Color Photographs of Artworks, Natural Scenes and Objects2012
- Spatial-temporal Video Quality Metric based on an estimation of QoE2011
- A survey of image quality measures2009
- A New Spatial Hue Angle Metric for Perceptual Image DifferenceLecture Notes in Computer Science, 2009
- Framework for Applying Full Reference Digital Image Quality Measures to Printed Images2009
- A top down description of S‐CIELAB and CIEDE2000Color Research & Application, 2003
- Image Kernels2002