شهاب خواجه غياثي

The detection of the inner and outer plexiform layers in retinal OCT scans is of great importance to ophthalmologists, as these layers play a key role in transmitting visual information. Changes in these layers can be indicative of eye diseases such as macular degeneration and diabetic retinopathy. Moreover, the plexiform layers are directly connected to nerve cells, and alterations in these layers can lead to neural damage, as seen in conditions like glaucoma. A precise examination of these layers not only aids in better diagnosis but also contributes to monitoring treatment progress and eva‎luating therapeutic effects. The main goal of this research is the automatic detection of these two layers using machine learning techniques and deep neural networks, without the need for direct specialist intervention. In this study, a convolutional neural network with a UNet architecture, which is highly effective and widely used in image processing and classification, is employed for the automatic detection of the layers. Additionally, to improve accuracy, the UNet architecture has been enhanced using the InceptionV3 model as a backbone.

يادگيري ماشين , پردازش تصوير , شبكه عصبي كانولوشنال , تصاوير OCT , لايه هاي پلكسيفرم

Machine Learning , Convolutional Neural Networks , OCT Images , Image Processing , Plexiform layers

Shahab Khajeh Ghiassi

Dr. Rahman Farnoosh