Noise Removal Using Histogram Equalized Based Contrast Masking For Image Quality Assessment

摘要

Image becomes the salient part, namely, watermarking, image enhancement, image compression, and etc. Similarly, the assessing the quality of the image also becomes an essential work of the users. In the last years, there are several advancements have been developed in analyzing the quality of the image. Considerably, Region-of-Interest (ROI) is used to assess the quality. ROI localization is a labor intensive process that takes multiple passes of sliding-window in search of proper ROI. Screen Content Image (SCI) comprises with picture regions and computer generated textual or graphical content. These are organized with statistical properties which lead to various behaviors. The SCI compression performance gets improved by the perceptual screen content coding scheme. Fetal US Image Quality Assessment (FUIQA) uses localization and classification to reduce the errors in the scanned images and enhances the quality of the image. Blind Image Quality Assessment (BIQA) predicts the quality of an image by showing the training data in the form of Discriminable image pairs (DIP) and then using the opinion-unaware BIQA model using RankNet Algorithms. This paper proposes a Noise Removal using Histogram Equalized based Contrast Masking scheme that reduced the time taken in ROI localization. The quality of the images is described using the features. ROI localization is performed using the Masking model which identifies masking and luminance value in parallel manner. The noise present in the image can be detected using the Finite Band Neighborhood Contrast measure.