CNN model for retinal image processing

The medical technology company, Aurteen Inc. from Calgary, Canada, with specialty in development of software for CAD of preterm retinal disorders and assessment of signs of severe diabetic retinopathy has engaged TROUT in a new project. Similar to other activities from TROUT, we again follow a non-invasive approach applying machine learning tools.


Building on past projects featuring ANN (artificial neural networks) and decision trees such as NIVAB and VITAB, TROUT will now generate a CNN (convolutional neural network) model for the recognition of blood vessels and the optic nerve head in retinal images. The next step will be the detection of suspicious and possibly pathological signs of the disease.

The picture on the right shows areal highlighting of a retina.