Optic disc detection in retinal fundus images using gravitational law-based edge detection

Area of Science:

• Ophthalmology
• Medical Imaging
• Computer Vision

Background:

• Diabetic retinopathy is a leading cause of global vision loss.
• Early detection and treatment are vital to prevent vision impairment.
• Accurate identification of the optic disc in retinal fundus images is a critical prerequisite for detecting diabetic retinopathy and segmenting blood vessels.

Purpose of the Study:

• To develop a simple, deterministic, and time-efficient approach for optic disc detection in retinal fundus images.
• To enhance the accuracy and reduce missed detections of the optic disc.

Main Methods:

• Adaptation of a gravitational law-inspired edge detection algorithm.
• Introduction of novel pre- and post-detection processing steps.
• Implementation of a candidate selection technique to minimize missed optic discs.

Main Results:

• Achieved high detection rates: 100% (DRIVE), 97.75% (DiaRet), 92.90% (DMED), and 95% (STARE).
• Demonstrated superior performance compared to existing optic disc detection methods.
• Exhibited a significantly fast average running time of 0.40 seconds per image.

Conclusions:

• The proposed method offers an accurate and highly efficient solution for optic disc detection.
• This approach can aid in the early diagnosis of diabetic retinopathy and improve automated analysis of retinal images.
• The speed and accuracy make it suitable for clinical applications requiring rapid image processing.