Pattern recognition in dermatology using k-nearest neighbors algorithms (k-NN)

Deep learning algorithms for reading MRI scans.​

Designing convolutional neural networks to analyze ECG strips.

Developing deep learning algorithms for diabetic retinopathy.​

Designing convolutional neural networks to analyze echocardiograms.

Predicting heart disease risk using machine algorithmic processing of electronic health records

We develop technologies involving creation of human like brainpower, technologies that can plan , perceive, reason, learn and process Natural Language.

MedWhat applies data science techniques to healthcare data stored in 2D medical images, 3D medical images, electronic health records, and wearable devices.

1. We can analyze data from the NIH and Hospitals Chest X-ray dataset and train a CNN to classify a given chest X-ray for the presence or absence of pneumonia. We can write an FDA 501(k) validation plan to clear the software for being used as a medical device.
2. We can create an algorithm that will help clinicians assess hippocampal volume in an automated way and integrate this algorithm into a clinician's working environment. We can also write up a validation plan that would help collect clinical evidence of the algorithm performance, similar to that required by regulatory authorities.
3. Our data scientists can help a pharmaceutical company that has created a diabetes drug that is ready for clinical testing. We build multi-variable regression models to predict the estimated hospitalization time for a patient in order to help select patients for your study.
4. Our team can build an algorithm which combines information from the IMU and PPG sensors that can estimate the wearer’s pulse rate in the presence of motion. Evaluate algorithm performance and iterate on design until the desired accuracy is achieved.