This software is designed for diagnosing retinal diseases based on OCT images and is capable of distinguishing between the following five classes: Age-related Macular Degeneration (AMD), Central Serous Retinopathy (CSR), Macular Hole (MH), Diabetic Retinopathy (DR), and Normal (NO).
In this program, images are processed at various scales using appropriate filters, and their features are extracted. The OCTID dataset has been used for training and evaluation. This dataset contains 572 images with dimensions of 512*750 across the five mentioned classes. For training the network within the program, 80% of the data was utilized, achieving an accuracy of 79.6%. Details of the detection method are outlined in the following paper. Please refer to this paper when using the software:
For testing, by selecting an OCT image with the specified dimensions, the software will display the detection result.
Download the software : Software Files
Image Selection: The mentioned dataset is accessible via the following link: