Hi, I'm Chen Liu (刘晨).
Self-driven, quick starter, passionate researcher with a curious mind who enjoys solving challenging and interesting problems regarding AI for healthcare.
Acknowledgements :Many thanks to Varad Bhogayata for kindly providing this cool website template.
About
I am currently a Research Assistant at the Smart Health Center at Hong Kong Polytechnic University, where I have the privilege of working closely with Dr. Wenfang Yao, Prof. Kejing Yin, and Prof. Harry Qin on advancing multi-modal learning applications in healthcare. My research journey began with a Bachelor's degree in Software Engineering from South China University of Technology.
I am also honored to be an intern at the Computer Cognition, Vision, and Learning (CCVL@JHU) Lab at Johns Hopkins University, where I work with Dr. Zongwei Zhou and Prof. Alan L. Yuille on exploring the application of generative AI in healthcare.
My passion lies in Multi-modality Learning and Medical Image Analysis. Recently, I have been particularly interested in modeling the time-varying dynamics of image sequences and irregular time-series data to enhance our understanding of disease progression.
Outside of research, I am an enthusiastic runner and a dedicated volleyball player. I love problem-solving and coding, always striving to bring my best effort to every project.
Looking for a PhD opportunity to work in a challenging position combining my skills in Machine Learning, Programming and Data analyses, which provides interesting experiences and personal growth.
Education
South China University of Technology
Guangzhou, China
Degree: Bachelor of Engineering in Software Engineering
GPA: 3.85/4.0
- Linear Algebra and Analytic Geometry (98)
- Probability and Mathematical Statistics (97)
- Database Systems (95)
- Operating System (94)
- Machine Learning (95)
Relevant Courseworks:
Publications
Experience
- Generated personalized CXR images using latent diffusion models, integrating asynchronous and heterogeneous CXR and EHR data for anatomical and disease progression information.
- Developed a multi-task EHR encoder to extract relevant imaging and disease progression data for dynamic image generation.
- Conducted extensive experiments and qualitative analyses to show that our model could generate high quality CXR images and outperform existing methods in multi-modal clinical.
- Key Words: Multi-modal Fusion, Diffusion Model, PyTorch
- Analyzed hard mode of AI segmentation in pancreatic tumors, identifying patterns for more realistic synthesis.
- Developed synthetic models for pancreatic tumors with varied characteristics using conditional diffusion models. Enhanced the diagnosis performance of different pancreatic tumor types on Current AI model.
- Participated in the construction of a large, publicly accessible lesion dataset with per-voxel annotations for 10,136 CT scans, including 60,038 lesions across six organs.
- Key words: Synthetic Tumor, Tumor Segmentation, Benchmarch Building
- Extracted semantic masks using MedSAM, which was integrated with input data as prior knowledge to facilitate the report generation.
- Transformed LaBERT from non-autoregressive to autoregressive, optimizing the inference method with beam search for comparable results.
- Developed a clinical loss function for image classification to improve finding awareness and proposed a method for extracting topic-related finding knowledge from pre-trained report models.
- Key Words: Radiology Report Generation, Controllable Image Captioning
- Developed a near-infrared retro-reflective patch for rapid normal vector measurement with optical navigation systems.
- Created an accurate robot arm positioning algorithm, integrating robot, marker, and optical tracking coordinates for precise injection angle control.
- Integrated coordinate acquisition, hand-eye calibration, and automatic injection into a C++/QT-based robot control platform.
- Key Words: medical robot control, hand-eye calibration
Skills
- Programming Languages: Python, C++, QT, Go, R
- Technologies: Artificial Neural Networks/Machine Learning (PyTorch, Sklearn, NumPy, Pandas), Data Processing, SQL,
- English proficiency: IELTS 7.0
- Tools: Git, Latex