Diagnosis of Purtscher Retinopathy from Spectral-Domain OCT Images Using Panoptic Segmentation Technique

Purtscher Retinopathy (PR) is a severe vision-threatening disorder that is frequently connected with trauma, severe pancreatitis, or connective tissue diseases.Conventional diagnostic techniques depend on fluorescein angiography and clinical examination, both of which can be intrusive and may not always yield conclusive results.Spectral-Domain Optical Coherence Tomography (SD-OCT) is a non-invasive imaging technique that produces high-resolution cross-sectional images of the retina and allows for accurate evaluation of retinal layers and illnesses.The proposed method implements a panoptic segmentation algorithm to detect PR using SD-OCT image data in an unsupervised manner.The Panoptic Feature Pyramid Network (FPN) detects and categorizes clinical indications of PR using instance and semantic segmentation.To detect additional challenging factors, a deep Convolutional Neural Network (CNN) is combined with encoder-decoder structures and instance segmentation networks.When it came to recognizing and classifying retinal disorders connected to PR, the proposed approach showed excellent accuracy.The performance of the proposed method was assessed using quantitative measures such as the Intersection over Union (IoU), Dice coefficient, pixel accuracy, precision, recall, and F1 score.This approach offers a non-invasive and efficient tool for the early detection of PR, allowing for timely management and better patient outcomes.