Image Interpolation with Edge Enhancement

概要

論文の詳細を見る
Interpolation is often used in image processing. Edges are blurred when images are processed with linear interpolations. In this paper, an algorithm for image interpolation with edge enhancement is discussed. To make detail edges in images more perceptible, images are expanded with spline interpolation. Intensity of all newly inserted pixels is determined based on the intensity distribution around it. After interpolation, edges are separated far away and it is easy to do the edge detection in interpolated images. Unlike other edge operators that convolute a smoothing function to images directly, the algorithm discussed in this paper smoothes the image by interpolation in which the image is expanded and smoothed, and thus detailed information remains. A unified linear-time algorithm for distance transform is used to deal with the calculation of Euclidean distance. So-called edge area (EA) is determined which is located around edges. Erosion is taken place in EA to determine the erode source intensity (ESI). Based upon distance. and ESI, intensity of every pixel is adjusted. Edge enhancement is performed repeatedly. By repeating the edge enhancement, edges in images get clearer.
一般社団法人情報処理学会の論文
2001-05-18