An Improvement of Edge-Adaptive Image Scaling Algorithm Based on Sobel Operator

In this paper, an improvement of fast image scaling algorithm, which produces image with well-preserved edge, is proposed. The edge of the original image is detected by using four Sobel spatial operators. And the direction of edge is then measured by using two newly proposed edge detectors employing two threshold values. Based on these, pixels are interpolated adaptively, where edge pixels are interpolated according to the edge direction and other pixels are interpolated by using bilinear interpolation. Experiment results show that the proposed algorithm produces images with higher peak signal-to-noise ratio (PSNR), higher structure similarity (SSIM) index, and better preserved edge than bilinear, bicubic and traditional fast edge-adaptive interpolations. When images have obvious edges, the scaled images are also subjectively looks better.