Seokhwan KO

Wilmington, MA 01887 ยท sko05@cs.tufts.edu

"I would rather fail in a cause that will ultimately triumph than to triumph in a cause that will ultimately fail."

- Woodrow Wilson

Experience

Deep Learning Researcher

Department of Pathology, Kyungpook National University Medical Center

-Designed a novel architecture integrating attention mechanisms and clustering methods to enhance interpretability and discriminate abnormalities in thyroid WSIs.

-Designed the 2-step cascaded architecture based on CNN and CRNN(CNN+RNN) to predict TERT promoter mutational status in thyroid cancer.

-Built 3D models to identify the correlation between appendicitis diagnoses from non-contrast CT scan and enhanced CT scan.

Apr 2021 - Present

Computer Vision Engineer

CAIDE Systems

-Developed a cascaded deep learning architecture, considering diverse window values to enhance the sensitivity of predicting hemorrhages in brain CT images.

-Designed novel deep neural networks by combining the VGG16 architecture with the Unet concept and modified pooling layers based on semantic segmentation method.

-Developed back-end engines for the deep learning medical platform to integrate in front-end UI/UX.

January 2017 - May 2021

Undergraduate Teaching Assistant

University of Massachusetts Lowell

Graded submitted materials and provided feedback with guidance for over a year in two classes of computer science department; COMP.1000 Media Computing, COMP.2030 Computer Org & Assembly Language.

September 2015 - December 2016

Skills

Programming Languages
  • Python, JAVA, C/C++
    Machine Learning
    • Caffe, PyTorch, Tensorflow
    • OpenCV, NumPy, PIL, SciPy, NiBabel, ITK
      Web Development
      • PHP, HTML, CSS, MySQL, Flask, Redis
      • Javascript, jQuery, JSON
        Development Environment
        • Linux(Ubuntu), Docker, AWS(EC2)

        Publication

        Deep Learning Prediction of TERT Promoter Mutation Status in Thyroid Cancer Using Histologic Images

        Medicina 2023
        March 2023

        Lesion Conditional Image Generation for Improved Segmentation of Intracranial Hemorrhage from CT Images

        arXiv preprint arXiv:2003.13868
        March 2020

        Improving Sensitivity on Identification and Delineation of Intracranial Hemorrhage Lesion Using Cascaded Deep Learning Models

        Journal of digital imaging, 32(3), 450-461
        January 2019

        Education

        University of Massachusetts Lowell

        Bachelor of Science
        Computer Science
        September 2013 - May 2017

        Interests

        On a weekend, I enjoy most of my time being outdoors. In the spring and summer season, I play baseball. And I also enjoy motorcycle riding.

        In my spare time, I get into music producing. I believe musical sensitivity creates proper motivation to those who pursue an optimistic life. And I spend my extra free time exploring and studying the latest technology advancements in the AI development world.