Tomoe Hagio
Development
Michigan, United States
Skills
Machine Learning (ML)
About
TOMOE HAGIO's skills align with IT R&D Professionals (Information and Communication Technology). TOMOE also has skills associated with Programmers (Information and Communication Technology). TOMOE HAGIO has 15 years of work experience.
View more
Work Experience
Research Scientist
INVIA Medical Imaging Solutions
January 2018 - December 2023
- Led deep learning projects and taught deep learning (including how to implement one) to other researchers using jupyter notebook Wrote python scripts and modules, and contributed to existing python packages (OOP) for data processing, deep learning algorithms, and training/validation routines Wrote R scripts and functions for data processing and statistical analysis Collected and cleaned large medical dataset (medical images and health record) using SQL and python (sqlite3, re, pandas packages) to prepare for deep learning projects Developed deep learning algorithms (includes implementing and modifying several generative models and optimizing hyperparameters) for SPECT images that resulted in publications in peer-reviewed journals Evaluated deep learning algorithms using linear regression, ROC analysis, and survival analysis for diagnostic and prognostic performance Deployed a generative deep learning algorithm and implemented other algorithms (image reconstruction, image processing) into an existing software written in C++ (GitHub team for version control) Optimized existing software written in C++ to improve localization of the heart Was invited to write an editorial for deep learning in a journal Parallelized data processing using multiple GPUs/CPUs Built computer and installed necessary tools (CUDA and other libraries) for developing deep learning algorithms on it Developed metrics for evaluating the quantitative parameter derived from PET images, often used in cardiac diagnosis, which led to a conference abstract and presentation
ORISE Research Fellow
United States Food and Drug Administration, Center for Devices and Radiological Health
January 2016 - December 2018
- Performed statistical assessment (regression analysis, etc.) of motion effect on quantitative coronary artery CT imaging using motion phantom Developed data visualization tool for complex datasets through MATLAB, R, and Python Implemented and evaluated performance of maximum likelihood-based statistical algorithm in Python and MATLAB Designed tissue-equivalent phantoms for dual energy CT imaging usage via simulation and analytical modeling of material behavior in dual energy CT Manufactured tissue-equivalent physical phantoms for validation Modeled 3D geometry and generated STL files for 3D printing Gained in-depth knowledge of statistical analysis in evaluating clinical software or devices Earned the Office of Science and Engineering Labs Quarterly Recognition Award in 2018 Sanofi Pharmaceutical R&D, Oro Valley, AZ Intern 2014 and 2016 Built an image analysis software and graphical interface to be used by chemists and biologists for a high-throughput analysis of drug effects through object-oriented MATLAB. This interface included semi-automated image segmentation and classification algorithms for microscopic video of a moving heart Developed a cost-effective drug-screening pipeline involving optimized high-speed microscopic imaging system (testing drug efficacy with different concentration, and signal/video processing through software I developed) Identified drug responses by conducting image and statistical analysis including feature extractions and regression analysis
The University of Arizona, MRI Research Lab, Department of Medical Imaging
January 2013 - December 2013
- Fuzzy C-means clustering applied to MRI data sets, in Mathematics 574M (Statistical Machine Learning).
The University of Arizona, MRI Research Lab, Department of Medical Imaging
January 2012 - December 2012
- Monte-Carlo simulation of anomalous diffusion, in Applied Math/Physiology 572 (Mathematical Modeling of Biological/Physiological System).
Graduate Research Associate
The University of Arizona, MRI Research Lab, Department of Medical Imaging
January 2009 - December 2016
- Facilitated interdisciplinary research (including clinical studies) and completed collaborative projects with radiologists and scientists with different disciplines Processed MRI signals and reconstructed MRI images using our custom reconstruction algorithm for undersampled data Developed machine learning algorithm (unsupervised learning) for feature extraction and characterization Optimized dimensionality reduction (unsupervised learning, e.g. PCA, SVD)-based image reconstruction algorithm to reconstruct sparse dataset Developed image segmentation algorithm for breast MRI images in collaboration with the signal and image laboratory Developed graphic user interfaces (GUI) to be utilized by radiologists and technicians for the automation of MRI image reconstruction, statistical analysis, results reporting, and automated MRI image segmentation Developed image analysis tools including automated image segmentation and image registration algorithms for various organs as well as algorithm packaging into a GUI Applied non-invasive MRI biomarkers for various disease diagnosis and prognosis predictions without the use of contrast agent, using MRI physics-based signal models and quantifying parameters from MRI images Modified GE pulse sequence utilizing EPIC (custom C language) programming Completed numerous multidisciplinary projects which involved collecting, processing (reconstructing), and analyzing MRI data, while optimizing clinical protocols and managing MRI data set for clinical studies and presenting reports of the findings to MRI researchers and clinicians Conducted repeatability and reproducibility studies for the image acquisition technique and quantification algorithms in computational simulation, phantoms and in vivo (healthy subjects and patients)
Education
University of Arizona
Doctor of Philosophy
University of Arizona
Master of Science in Biomedical Engineering
