Image resolution enhancement method using multiple interpolations

摘要

Image interpolation is a method of determining the value of new pixel coordinate in the process of image scaling. As a large-capacity image contents, interpolation algorithm is required to generate enhanced result image as high-speed. In this paper, fast multiple mixed image interpolation for image resolution enhancement is proposed. The proposed method estimates expected 12 errors from four subimage of input image and generate result image interpolation of combine expected error and input image. The experimental results demonstrate that PSNR increases maximum to a value of 2 dB, SSIM increases maximum to a value of 0.05, and the subjective quality is superior to any other compared methods. Moreover, it is known by algorithm running time comparison that the proposed method has better faster over three times when compared with conventional methods. The proposed method can be useful for application on image resolution enhancement.