基于阈值控制的边缘自适应快速图像插值算法

在处理视频、网络信号时，要求较低的算法复杂度和运算量。该文通过阈值控制方式分类像素点区域，采用双通道达到快速插值的目的。利用与边缘像素点相邻的6个降采样像素局部结构多方向信息特征，估计高分辨率像素值。实验结果证明，该算法有效地保持了边缘特征，减少了CPU处理插值的时间，获得了较高的分辨率图像，并满足快速处理信号的要求。

Fast Self-adaptive Edge-oriented Interpolation Based on Threshold Method

Lower computational complexity of the algorithm and better subjective quality are important in video and network signal processing. The interpolation of pixel is accomplished by double-channel method. To estimate relative information at high resolution, the algorithm analyzes the multiple directions relativity of local neighbors and makes full use of six down-sampled neighbors involved in the low resolution image, Experimental results demonstrate that algorithm not only decreases the computational complexity but also substantially improves the visual quality of the interpolated images.