The paper sets forth an improved edge-directed image interpolation algorithm with low time complexity. The algorithm partitions images into homogeneous and edge areas by setting the preset threshold value based on the local structure characteristics. Specified algorithms are assigned to interpolate each classified areas respectively. The proposed method implements strategy in three steps to interpolate after setting the preset threshold value. In this way, it can achieve the goals of real-time interpolation and good subjective quality. Furthermore, the interpolated images have much more explicit edge regions and better visual effects using our proposed method than that of using other algorithms. Experimental results demonstrate that the method proposed by the authors is high-performed in image interpolation.
Low-Cost Implementation of Bilinear and Bicubic Image Interpolation for Real-Time Image Super-Resolution