ConHyp-Seg- Deep Neural Network Based Conjunctivia Hyperemia Segmentation with Mask Categorization Grading Networks

Ocular redness is a key indicator of inflammation and can signify the severity and progression of various illnesses. The conjunctiva, a crucial protective layer of the eye, plays a vital role in maintaining ocular health, and its impairment can lead to vision-related complications. Conjunctival hyperemia, often associated with viral or bacterial infections, results from dilating conjunctival blood vessels, causing redness and swelling. Accurate assessment of this condition requires precise segmentation and grading of affected regions. This study presents ConHyp-Seg, a deep neural network-based framework for conjunctival segmentation from eye images, and Mask Categorization Grading Networks (MCGN), a complementary system for classifying hyperemia into four severity grades. Integrating these two models offers a robust AI-powered solution for diagnosing and grading conjunctival hyperemia, facilitating early and effective intervention. The proposed system achieves exceptional performance, with 98.2% accuracy, 97.8% precision, 98% recall, and an F1-score of 0.98. These results underscore its potential as a reliable diagnostic tool to assist ophthalmologists in making precise, objective assessments, ultimately enhancing patient care and disease management.