DefinePK

Abstract

Image compression is a very important issue for many applications in the field of visual communications. The purpose of image compression is to reduce the storage and transmission costs while maintaining the image quality. Nowadays, image previewing on a mobile device is a practice that takes a great implementation of everyday live. However, the image compression with very low bit rates that is typically used on mobile platforms, usually may introduce visible compression artifacts, which is referred to blocky artifacts. This displaced blocky artifacts, although visually noticeable and annoying, is particularly difficult for automated detection because its location is a priori unknown, and its appearance might be easily mistaken for some real edges or fine details in the image. Therefore, the detection of the blocky artifacts is important to ensure the de blocking process is performed on the blocky artifacts only, not on real edges or fine details of the image. This paper proposes the development of a technique which aims in blocky artifacts detection for compressed digital image. The detection of blocky artifacts presented in this paper is performed in two parts that are vertical and horizontal detections utilizing proposed detection algorithms. The effectiveness of to detect the blocky artifacts detection without including the image edges and fine details is depending on the threshold value setting in the vertical and horizontal detections algorithms. After both detections are completed, then the last step is to combine both edge maps into a new image which includes the blocky artifacts at the boundary between two different luminance gradients. The algorithm is developed in MATLAB software. The analysis for the result is made based on qualitative observation. For blocky artifacts detection, the proposed technique has achieved its objectives in detecting blocky artifacts at the boundary between two different luminance gradients. Thresholding process separated the unwanted image fine details and edges, providing an output image a view of clearer blocky artifacts existing in the compressed image. It is also found that the proposed technique could detect blocky artifacts more effectively (without including image fine details and edges) in comparison to conventional techniques.