NEW GRADIENT EDGE DETECTION METHOD BASED ON CAPUTO $k$-FRACTIONAL DERIVATIVE

Résumé

In this paper, we propose a new gradient edge detection  method based on the Caputo $k$-fractional derivative. In this method,  we construct a new $ 3×3 $ fractional order mask to obtain the gradient components. The main property of this technique is that it has a double memory effect in detecting edges. The experimental results for some natural and medical test images show that our proposed edge detection  method outperforms both Caputo and Caputo-Fabrizio fractional methods in terms of both visual perception and peak signal to noise ratio.