Image enlargement using Adaptive Manipulation Interpolation Kernel based on local image data
I Komang Somawirata, Keiichi Uchimura, Gou Koutaki · 2012
This paper proposed Adaptive Manipulation Interpolation Kernel (AMIK) based on local image data. The adaptive interpolation kernel used closed loop system. The previous output result fo'(i,j) also be used as reference for the next interpolation calculation. Closed loop applied in the system, if the interpolation system performed between two pixels for horizontal or vertical directions. Furthermore, fo'(i,j) filtered by Three Mean Median Filter (TMMF). That method used to increase image quality. TMMF is the average value of three median nearest pixels in the 3×3 window size. The test results shown good results compared to the Bilinear, Bicubic, Lanczos 2, Lanczos 3 and Pyramid step. PSNR is the quantitative analysis, which used to measure image performance. Our method has the highest PSNR value in the gray and RGB image, which compared to the comparison methods.