Fractional Fourier Transform based Riesz fractional derivative approach for edge detection and its application in image enhancement

• Published in: Signal processing Vol. 180; p. 107852
• Main Authors: Kaur, Kanwarpreet, Jindal, Neeru, Singh, Kulbir
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.03.2021
• Subjects: Edge preservation index; Figure of merit; Fractional Fourier Transform; Riesz fractional derivative; Fractional Fourier Transform; Edge preservation index; Riesz fractional derivative; Figure of merit
• ISSN: 0165-1684; 1872-7557
• DOI: 10.1016/j.sigpro.2020.107852

•Mathematical framework for Riesz Fractional Derivative (RFD) is obtained using Lagrange, Newton, and Aitken Interpolation method.•RFD in Fractional Fourier Transform (FrFT) domain is proposed for edge detection.•Edges obtained with proposed approach are used for the application of image enhancement.•Uncontrolled features such as different illumination conditions, Poisson noise, and JPEG compression artifacts are considered to confirm the effectiveness of proposed approach. Edge detection plays a key role in detecting the boundaries of an object in the image to improve the quality of image edges. The edge detection techniques based on integer order derivatives are more prevalent in literature. But these techniques have the limitation of providing thicker edges and being sensitive to noise. Therefore, in order to deal with this issue, an edge detection technique based on Riesz fractional derivative (RFD) in Fractional Fourier Transform (FrFT) domain is proposed in this paper. The RFD mask used for edge detection is obtained by using various interpolation methods. The selection of mask size is done on the basis of Figure of Merit (FOM) and Edge Preservation Index (EPI). The edges obtained with proposed approach in FrFT domain are further used for image enhancement. The effectiveness of proposed approach is confirmed using test images from different standard datasets by considering performance metrics such as FOM, EPI, Information Entropy, Average Gradient, Edge Intensity, etc. Further validation of the proposed approach is done by considering various uncontrolled features such as Poisson noise, different illumination conditions, JPEG compression artifacts, etc. Experimental results show that proposed technique outperforms the existing techniques by providing improvement in various performance parameters.