Varun Shenoy

During the month of August in 2018, I developed a web tool for bone x-ray annotation as well as the backend infrastructure for image masking and cropping. It will fit into the full application pipeline which includes a machine learning system to clasify bone features and tumors using the masked regions.

Varun Shenoy and other authors from Stanford University (2018)

Publication pending

I spent the summer of 2018 at the UC Berkeley Artificial Intelligence Research (BAIR) Lab building 3D and 2D renditions of segmentation architectures for automated brain tumor semantic segmentation in MRI scans. I set up a 3D Unet for training with patch normalization and percentile thresholding. I developed a variety of different 2D segmentation algorithms to evaluate axial brain slices one at a time, including CycleGAN generated data. Finally, I implemented a masking strategy that leverages whole tumor segmentation maps to bolster dice scores on enhancing tumor and tumor core segmentations.

Amir Gholami, Shashank Subramanian, Varun Shenoy, Naveen Himthani, Xiangyu Yue, Sicheng Zhao, Peter Jin, George Biros, Kurt Keutzer (2018)

Accepted to Springer Lecture Notes in Computer Science (LNCS)

The incidence of postoperative wound infections after lower extremity bypass can be as high as 10%-20%. An automated method of diagnosing wound complications would serve to limit the expense of time and money from hospitals, doctors, insurers, and patients. The algorithmic classification of wound images, due to variability in the appearance of wound sites, is a challenge. Deep convolutional neural networks (CNNs), a subgroup of artificial neural networks that exhibit great promise in the analysis of of visual imagery, may be leveraged to categorize surgical site wounds. We present Deepwound, a multi-label CNN trained to classify wound images with image pixels and labels as the sole inputs.

Varun Shenoy, Elizabeth Foster, Lauren Aalami, Bakar Majeed, Oliver Aalami (2018)

Accepted to 2018 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

Plenary Session Presentation at the 2018 Surgical Infection Society Annual Meeting (Top 10 Papers)

Patent pending under US 62/670,970

Biometric measurements captured from medical devices, such as blood pressure gauges, glucose monitors, and weighing scales, are essential to tracking a patient's health. Trends in these measurements can accurately track diabetes, cardiovascular issues, and assist medication management for patients. Currently, patients record their results and date of measurement in a physical notebook. It may be weeks before a doctor sees a patient’s records and can assess the health of the patient. This research presents a mobile application that enables users to capture medical monitor data and send it to their doctor swiftly using HealthKit. A key contribution of this paper is a robust engine that can recognized digits from medical monitors with an accuracy of 98.2%.

Varun Shenoy, Oliver Aalami (2017)

Oral Presentation at the 2017 American Medical Informatics Association Annual Symposium