Abstract:

Digital contents, such as computer images in any form and type, are usually large and require substantial amount of server spaces to store it permanently. The increasing rate at which the volume of digitized images occupies an organization’s server is overwhelming, and hence, may potentially use up all the available storage spaces quickly. Therefore, there is a need for an image interpolation technique to be developed to solve the problem and help to improve content management in any organization that needs to keep large amount of digital images. However, there have been many image magnification techniques introduced in the past, but only until recent decade has seen new proposed methods being developed to deal with color image interpolation. Most of the proposed algorithms seem to work well to solve aliasing effect on computer graphics; however, many of the proposed methods do not consider minimizing edge blur effect on the magnified image. This study proposed a hybridized interpolation method that combines a bilinear interpolation method with an edge detection method. The proposed method consists of three stages. An image is replicated to two replicas. The first stage is to interpolate the first replica to four times its size using the bilinear interpolation. In the next stage, the second replica will be converted into 3-bit color palette image for easy edge detection work to be performed. The detected edges are maintained while other pixels are removed. The proposed Geometrical Pattern Classification approach will be conducted to the maintained pixels along the edges for touch-up work to be applied to color the pixels with correctly chosen colors in the positions where jaggies normally occur due to magnification process. This is also performed to minimize the smoothing effect caused by the bilinear algorithm taking the weighted average of the surrounding pixels. The final stage is to map the outline of the detected edges into the first replica to produce clear and smooth magnification result, which the proposed technique is expected to contribute in the research. The initial result was evaluated by conducting the satisfaction survey, which is based on these factors such as clarity, blur effect and sharpness of the magnified image.