System for providing artificial intelligence-based retinal disease diagnosis assistance service

A system for providing an artificial intelligence (AI)-based retinal disease diagnosis assistance service, the system comprising:

• a user terminal configured to transmit tomographic images of a retina captured by an optical coherence tomography (OCT) device, which uses a non-destructive examination method, and receive and output a retinal disease classification result;
• and
• a diagnosis assistance service provision server including a preprocessing unit configured to receive and preprocess the tomographic images, a diagnosis assistance unit configured to input the preprocessed tomographic images into a prebuilt AI algorithm, a visualization unit configured to visualize a retinal disease classified by the prebuilt AI algorithm, and a sharing unit configured to provide the retinal disease classified by the prebuilt AI algorithm to the user terminal and transmit a visualization result, wherein the prebuilt AI algorithm comprises:
• a slice-wise image (SWI) encoder configured to generate embeddings, which are summary information of the preprocessed tomographic images, based on the preprocessed tomographic images are input;
• a lesion-attentive fusion (LAF) layer configured to automatically detect preset lesion images among the preprocessed tomographic images, convert the detected lesion images into a one-dimensional (1D) array, and then transmit the 1D array;
• and
• a retinal disease classifier configured to classify the 1D array input from the LAF layer into a preset retinal disease class using a fully connected layer,
• wherein the SWI encoder generates the embeddings using a convolutional neural network (CNN)-based image classification model comprising a plurality of layers, each layer connected to every preceding layer of the plurality of layers, such that accuracy of the prebuilt AI algorithm is improved by reducing information loss, and improving gradient vanishing.