Authors: Manisha Jadhav; Yogesh Dandawate; Narayan Pisharoty
Addresses: Symbiosis International University, Pune, Maharashtra, India; Department of Electronics and Telecommunication Engineering, Marathwada Mitra Mandal's College of Engineering, Pune, Maharashtra, India ' Department of Electronics and Telecommunication Engineering, Vishwakarma Institute of Information Technology, Pune, Maharashtra, India ' Department of Electronics and Telecommunication Engineering, Symbiosis Institute of Technology, Symbiosis International University, Pune, Maharashtra, India
Abstract: Today's world has witnessed tremendous increase in use of multimedia and internet. This demands an image quality metric capable of evaluating image quality, accurately and automatically. Natural scenes are of excellent quality and all natural scenes exhibit similar statistical properties. Natural scene statistics is successfully used in image quality assessment which is based on the hypothesis that introduction of distortion in an image causes deviation from statistical properties. Amount of deviation in the statistical property of an image is found to be proportional to the amount of distortion. A neural network-based image quality metric needs such natural scene statistical feature to predict the image quality blindly. This paper presents a new feature for colour image quality assessment that is extracted after decomposing given image into different frequency bands of hue plane. In future, this feature will be used in a classifier to evaluate colour image quality.
Keywords: colour images; image quality; image distortion; human visual system; HVS; natural scene statistics; NSS; Gaussian distribution; HSV colourspace; quality assessment; Laplacian pyramid decomposition.
International Journal of Computational Vision and Robotics, 2015 Vol.5 No.4, pp.407 - 421
Received: 28 Aug 2014
Accepted: 06 Oct 2014
Published online: 20 Aug 2015 *