Two-Stage Interpolation Algorithm Based on Fuzzy Logics and Edges Features for Image Zooming

This work presents an innovative two-stage interpolation algorithm for image resolution enhancement and zooming applications. The desired high-resolution images are obtained via two interpolative stages. In the first stage, aligned pixels are first estimated using a fuzzy inference system, whose critical parameters are optimized by particle swarm intelligence. In the second stage, interior pixels are then restored by utilizing the edge properties of nearby pixels. From experimental results, numerical comparison confirms the superiority of the proposed interpolation algorithm over other existing methods. Furthermore, visual illustrations including zoomed parts and error maps demonstrate the significant improvement of the proposed method, particularly in the regions that contain many local edges and sharp details.