Improved image processing techniques for optic disc segmentation in retinal fundus images

•Automatic optic disc localization in fundus images for detection of retinal diseases using pixel density calculation method.•Optic disc segmentation through improved circular hough peak value selection and red channel superpixel segmentation.•Retinal image databases viz. HRF, DRISHTI-GS1, DRIONS, DRIVE, ONHSD, CHASE-DB1,INSPIRE and MESSIDOR are used for this work.•The proposed method yield better accuracy in optic disc localization and segmentation when compared to existing methods.•Achieved an accuracy of 99.93% for disc localization and 99.5% for disc segmentation. Glaucoma is one of the leading causes of blindness in the world and is projected to affect over 79.6 million people globally. Recently, automated computer aided systems are used in disease detection and proven to be highly useful in assisting experts in the early diagnosis. Hence, automated optic disc segmentation through the intelligent system is very much helpful for the early detection of Glaucoma. This paper presents an improved image processing algorithm for retinal fundus images using region based Pixel density calculation method for optic disc localization and improved Circular Hough Transform with Hough Peak value selection and Red channel Superpixel segmentation for Optic disc segmentation. Optic disc segmentation has been applied on eight publically available databases, HRF, DRISHTI-GS1, DRIONS-DB, DRIVE, ONHSD, CHASE-DB1, INSPIRE and MESSIDOR and the accuracy of 99.73%, 99.31%, 99.37%, 99.38%, 99.64%, 99.20%, 99.31%, 99.72% and also specificity of 99.90%, 99.43%, 99.60%, 99.52%, 99.82%, 99.43%, 99.84%, 99.89% were obtained with less than 2 s computation time for an image. The result of the proposed technique shows that the system is highly competitive with the state-of-the-art and achieves better accuracy with fast execution time.