Three level automatic segmentation of optic disc using LAB color space contours and morphological operation
Shree Prakash,Jagadeesh Kakarla
DOI: https://doi.org/10.1002/ima.22895
IF: 2.177
2023-04-28
International Journal of Imaging Systems and Technology
Abstract:Fundus disorders like Glaucoma are a serious health risk that affects people all over the world, lowering their quality of life. Optic disc (OD) segmentation is a crucial step for automatic detection of glaucoma with fundus images. This paper presents an efficient and accurate OD segmentation methodology from eye fundus images. In the proposed methodology, fundus images are transformed from RGB (Red, Green, and Blue) color space to LAB (Luminance, A‐axis, B‐axis) color space. A new image is generated using the luminance, A‐axis component, and image information content. As a pre‐segmentation stage, the unsupervised approach extracts the region of interest containing OD. The adaptive threshold method, followed by morphological operations, extracts the contours of the OD region and removes the spur. Segmentation of OD is a challenging task due to similar intensity levels with neighborhood, vascular occlusion, and retinal atrophy. The proposed work overcomes the challenges by segmentation of the OD at three levels, which enhances the functionality at different illumination conditions. The features are extracted at various levels to determine whether the segmented image contains the OD or not. The efficiency of the proposed methodology demonstrates by applying it to a variety of fundus images from datasets, namely ISBI IDRiD, DRISHTI‐GS, and SYSU. Key performance metrics like Intersection over Union and Dice Coefficient are used to evaluate the performance of the proposed methodology. The proposed methodology achieves Intersection over Union of 0.996, 0.997, and 0.994 on ISBI IDRiD, DRISHTI‐GS, and SYSU datasets, respectively. Similarly, the proposed methodology obtains 0.96, 0.97, and 0.98 Dice Coefficient on the three datasets.
engineering, electrical & electronic,optics,imaging science & photographic technology