Application of fuzzy inference and active contour model for detection of fovea and its center in a fundus image

Abstract

Detecting the center of the fovea and its boundary in a human retinal image is a challenging task due to the irregularity of the avascular foveal region. Several attempts have been made in order to detect the boundary using fluorescein angiographic images. However, no attempt was made until now in detecting the boundary of fovea directly from fundus images. In this paper, a novel method for detecting the center of the fovea and its boundary in a fundus image is proposed. A fuzzy rule-based technique has been implemented along with a gradient vector flow (GVF)-based active contour technique in order to detect the boundary of the foveal avascular zone as well as its center. The proposed method exhibits to be promising in detecting fovea regions in wide range of images.