沿自适应图像缩放算法

为了找到正确的图像边沿去进行插值，提出了一种改进的边沿自适应图像缩放算法.计算图像中各像素处不同边沿方向的相关度，用模糊度参数对检测出的边沿相关度进行过滤处理，将相关度值最小的方向预选为该像素的边沿方向，生成整幅图像的边沿信息表.根据物体边沿信息的连续性，利用延续度参数优化边沿信息，得到最终使用的边沿信息表.对于图像中存在边沿的区域，插值沿边沿方向进行.对于图像中不存在边沿的区域，采用传统的插值方法进行插值.仿真结果表明,在人造图像和自然图像中，该算法都能有效消除传统插值方法在物体边缘的模糊和锯齿现象.

Edge adaptive image scaling algorithm

To find the correct edges of an image and interpolate with the edge, an improved edge adaptive image-scaling algorithm was proposed. The correlation of different edge direction was calculated at each pixel of the image. The ambiguity parameter was adopted to filter the detected correlation of edge information. The direction of the minimal correlation was set as the direction of the edge at the pixel. An edge information table of the whole image was built up. The continuity parameter was adopted to optimize the edge information according to the continuity of edge information. The last table of edge information was found. Interpolation was implemented along the edge direction in the region with edge. The interpolation was done with traditional interpolation method in the region without edge. Simulation results show that for natural images and synthetic images, the proposed algorithm effectively removed the sawtooth and blur at the edge caused by traditional interpolation method.