Author(s): PS Parsania and Dr. PV Virparia

Abstract: Images are extensively used in various applications on social media and over the Internet. To view these images on various display devices, image dimensions are changed according to display size. While adjusting an image it goes through the image scaling operation so that image can be properly viewed. To scale an image, one needs to use image scaling algorithm. Various algorithms can be used to determine the new value of the unknown pixel, all these algorithms perform interpolation amongst the nearby pixels in the original image. In this research work, eight image scaling algorithms are used that include Nearest neighbor, Bilinear, Bicubic, Cubic B-Spline, Catmull-Rom, Lanczos order 2, Lanczos order 3, and newly proposed algorithm. The computational time complexity is measured using a single computer and cluster computing for each of these image scaling algorithms and compared with the newly proposed algorithm. The developed algorithm is based on the efficient cubic B-spline function that enhances the image by smooth interpolation at the image edge. Subsequent images generated using various algorithms are then evaluated for their quality using PSNR and UIQI image quality assessment techniques.

DOI: 10.33545/27076636.2021.v2.i2a.95

Pages: 39-45 | Views: 248 | Downloads: 92

Download Full Article: Click Here