Document Type : Research Articles

Authors

• M Khanian ¹
• A. Davari ²

¹ Department of Mathematics, Khorasgan (Isfahan) Branch, Islamic Azad University, Isfahan, Iran.

² Department of Mathematics, Faculty of sciences, University of Isfahan, Isfahan, Iran.

Abstract

Image restoration has been an active research area. Dierent formulations are eective in high quality
recovery. Partial Dierential Equations (PDEs) have become an important tool in image processing
and analysis. One of the earliest models based on PDEs is Perona-Malik model that is a kind
of anisotropic diusion (ANDI) lter. Anisotropic diusion lter has become a valuable tool in
dierent elds of image processing specially denoising. This lter can remove noises without degrading
sharp details such as lines and edges. It is running by an iterative numerical method. Therefore, a
fundamental feature of anisotropic diusion procedure is the necessity to decide when to stop the
iterations. This paper proposes the modied stopping criterion that from the viewpoints of complexity
and speed is examined. Experiments show that it has acceptable speed without suering from the
problem of computational complexity.