No Reference Gradient Oriented Image Quality Assessment

Image quality assessment plays a ubiquitous role in the applications of image processing. As we are familiar, the digital images are most likely to subject wide range of distortions in the process of preprocessing, compression, storage, transmission, etc. Quality of images can be perceived by using IQA metrics. IQA metrics are of two types subjective image quality metrics that perform best only in the human visual system and second type of metrics is objective metrics, which are more appropriate than subjective metrics as they are more convenient and less time consuming. In spite of great advancement in the research area, no reference image quality assessment is still a challenging task because of the lack of the reference image. No reference image quality assessment evaluates the perceptual quality of the distorted image without its reference image. Existing BIQA methods mainly focused on the picture quality by breaking down the image insights in DCT, DWT domains. In this paper, we propose a novel no reference image quality assessment method using Laplacian of Gaussian features and Gradient Orientation. Previously gradient orientation is not much explored deeply as the main source of information for image quality assessment. In this paper, we apply the gradient orientation relative to all directions.