to help clinicians in the Real World.
To ensure that our models are robust and can tolerate the variety in real world data, we AUGMENT our training datasets through random transformations.
After the model has been trained on large datasets, we evaluate our models using data that have NOT been seen by the model before.
We deploy our models on our website using powerful TensorFlow.JS technology.
Our models run locally on your machine to ensure privacy and security of your data. Data is NOT sent to a server for inference.
We created a Tympanic Membrane image classifier based on Deep Learning that is capable of classifying given images into multiple pathologies.
We created a Thyroidectomy classifier based on Deep Learning that is capable of classifying based on set criteria.