Scott Trinkle

I am a data scientist with experience developing image processing, segmentation, and classification tools for image data. I completed my Ph.D. in Medical Physics at the University of Chicago in December 2021. My work there involved developing computer vision and image registration tools for validation studies between multiple preclinical imaging modalities including synchrotron x-ray microCT, diffusion MRI tractography, fluorescence microscopy, and electron microscopy. Now at Waters Corporation, I work as a Data Engineer developing Python web apps to perform automated tissue classification and interactive visualization of mass spectrometry imaging data. In my free time, I can usually be found watching something from the Criterion Channel, playing guitar, running, rock climbing, or baking bread.

Braingraphgeo

Python package for performing analysis on structural brain networks using random geometric surrogate graphs.

Fiberorient

Python package for extracting and visualizing the orientations of local structures in 3D imaging data using structure tensor analysis.

Nuclei Finder

Personal side-project for the segmentation of cell nuclei in multiple optical imaging modalities. Model built with TensorFlow using a UNET architecture, app built and deployed with Streamlit.

Waters Corporation

Data Engineer, January 2022 - Present

University of Chicago Committee on Medical Physics

Graduate Research Assistant, September 2016 - December 2021

University of Florida Advanced Laboratory for Radiation Dosimetry Studies

Undergraduate Research Assistant, January 2013 - May 2016

New preprint: Model-based vs. model-free myelin imaging

My latest manuscript uses model-free analysis of echo-planar spectroscopic imaging to show that fitting data to common biophysical models lowers myelin classification performance.

New position: Data Engineer at Waters Corporation

I have been hired as a Machine Learning Intern at Waters Corporation, where I will be developing python web apps for mass spectrometry imaging.

Ph.D Thesis defense

I successfully defended my thesis entitled "Multi-modal validation of MR microstructure imaging in the mouse brain.

New paper: Geometric bias in mouse brain networks from diffusion MRI

My new manuscript uses graph theory and optical tracer imaging to show that many properties of structural mouse brain networks measured with diffusion MRI can be largely explained through their spatial embedding alone, revealing geometric biases in diffusion tractography.

Paper highlight: Figure chosen for cover of Magnetic Resonance in Medicine

Our figure displaying spatial registration results between microCT and diffusion MRI is featured on the August 2021 cover.

Programming

Python (NumPy, Pandas, Scikit-learn, Matplotlib, Bokeh, Keras, TensorFlow, PyTorch), Bash, MATLAB, SQL, R, C++

Analysis

Computer vision, machine learning, image segmentation, network analysis

Tools

git, GNU Emacs, LaTeX, Jupyter, Docker, AWS