🐻💜

Junqi (Kuma) Liu

Vision-Language Model & Medical Image Researcher 🧠

Baltimore, MD 🇺🇸 • jliu452@uw.edu

My Learning Path

University of Copenhagen 🇩🇰 MSc IT & Cognition

Focused on NLP and Computational Cognitive Science.

NLP Cognitive Science
Shanghai Jiao Tong University 🇨🇳 BEng Artificial Intelligence

Focused on Machine Learning, Computer Vision, and Reinforcement Learning.

Machine Learning CV NLP

Research Adventures

Diffusion Model in CT Image Enhancement Johns Hopkins University
📍 CCVL Group | May 2025 - Present
  • Mission: Developed "SMILE," the anatomical aware image enhancement tool for medical image enhancement.
  • Achievement: 50% better FID in image quality, and 10% lift for F1-score in early tumor detection.
  • Publication: RSNA 2025 oral presentation !🎤 & Submitted to CVPR 2026 🚀.
🐙 Code 📜 Paper 🌍 Website 🤖 Model
Medical Imaging (ShapeKit) Johns Hopkins University
📍 CCVL Group | May 2025 - Present
  • Mission: Developed "ShapeKit," an efficient CT postprocessing pipeline.
  • Achievement: Lift average over 10% DSC score on organ segmentations.
  • Publication: MICCAI 2025 poster presentation !🎉
🐙 Code 📜 Paper
Zebrafish Tracking AI University of Copenhagen
📍 Kermen Lab (Neuroscience) | Oct 2024 - May 2025
  • Mission: Developed a Multi-Attention tracking model to detect Zebrafish biological movements. 🐟
  • Achievement: Over 80% sensitivity and 90% specificity.
  • Publication: Submitting a paper to PNAS journal 🚀.
🐙 Code
Explainable AI for Optimization Shanghai AI Lab & SJTU
📍 Yes Lab | Oct 2022 - Oct 2023
  • Mission: Innovative design for Bayesian Optimization, for multi-location geological perspective task.
  • Achievement: Lifted accuracy by 50%, and saved at least 90% time consumption! 🚀
🌍 website

Skills & Magic

Python 🐍 PyTorch 🔥 Tensorflow SQL Data Viz