Deep Learning Performance Accuracy

First application product of brain stroke diagnosis program (BSR) is using an AI model set that were trained using 6,000 stroke cases and more than 140,000 CT images. Based on 10 folder cross validation, BSR model achieved more than 90% of sensitivity accuracy on five sub types of hemorrhagic conditions.

Accuracy is significant.

Our core technology team invented smart data augmentation and improved overall accuracy of the model.

Our team never stops developing more efficient and accurate deep learning model for medical image classification and segmentation.

CAIDE Systems is still working on medical AI models to produce improved results.

Go Back