placeholder image
  • Contact Info

Kim, Geena Instructor

Positions

Research Areas research areas

Research

research overview

  • My research interest is in applications of artificial neural networks. Previously I worked on computer vision problems using neural networks for healthcare applications: 1) brain tumor substructure segmentation using neural network architectures on multimodal brain MRI data, and 2) prediction of blood pressure in the coronary arteries and STENT surgery recommendation using neural network and machine learning models on intravascular ultrasound images. I’m also interested in studying/emulating cognitive mechanism in brain. My recent projects include 1) learning association of vision and sound using generative adversarial network, 2) utilizing memory structure and curriculum learning for improving neural network training (currently applying to the brain tumor segmentation task), 3) how a visual memory can be represented as a graph and be used in a navigation problem using deep reinforcement learning.

keywords

  • machine learning, artificial intelligence, medical image

Background

Other Profiles