Ian Pan

Apply machine learning to various medical image analysis problems.

Automatic detection and segmentation of intracranial hemorrhage in head CT scans using 3D convolutional neural networks (CNNs).

Screening for pathological findings in chest radiographs using CNNs.

Bone age estimation from pediatric hand radiographs using CNNs.

Malignancy scoring of thyroid nodules viewed on ultrasound using CNNs.

Extracted and reviewed quantitative data from over 300 clinical trials and observational studies.

Proofread and edited data extractions performed by other data extractors.

Assembled all data into a central dataset for client use.

Contributed to systematic review examining efficacy of tympanostomy tubes in children with otitis media.

May 2015 - Aug 2015

Used data science to identify women at high risk of experiencing an adverse birth outcome.

Developed a machine learning pipeline to extract, transform, and load data from administrative databases that were used to train machine learning models.

Developed a web application to facilitate client and provider usage in the field.

Published methodology and findings in the American Journal of Public Health: Pan et al. (2017) Machine Learning for Social Services: A Study of Prenatal Case Management.

GitHub open source repository available here.