kazuhasasd/EyeDisease-Classifier

a advanced CNN + GNN combined hybrid eye disease classification model trained to identify common patterns in over 10+ types of eye diseases / infections.

GNN Model

the GNN model is a graphing neural network that in this case graphs out the special features picked up by the CNN model and looks for similar patterns, this is really important since there are alot of similar looking diseases that are not as easy to distinguish which can lead to the patient being given the wrong medication

CNN Model

the CNN model was trained on over 10k images of medical fundus photos of infected eyes it uses 2 extra max pool layers to obtain as much detail as possible to find more features to distinguish between the diseases it's been trained on