Image Resolution Conversion Based on an Edge-Adaptive Interpolation Kernel
Takashi Mori, Keisuke Kameyama, Yasuhiro Ohmiya, Jia Lee, Kazuo Toraichi · 2007
In this paper we propose a novel edge-adaptive interpolation method which is easy to implement and computationally efficient. Although various edge-adaptive image interpolation methods have been proposed in the past, none of them are comparable to classical interpolation methods such as Bi-linear interpolation or Bi-cubic interpolation concerning computational cost. Here we propose a novel interpolation method modifying the classical interpolation methods adaptive to improve the quality around the edges. The proposed method's flow is the following: 1) For all pixels, determine the local gradient defined by plane approximation of four neighbouring pixels. 2) Derive the edge-adaptive interpolation kernel by edge-directed interpolation axes. 3) Interpolate. From the results of experiments, the effectiveness of the proposed method concerning the computational cost and subjective image quality have been verified.