Han Liu

Title: Han Liu: Advancing Medical Imaging Analysis Through Machine Learning.

Introduction

Han Liu is an inventor based in Nashville, Tennessee, whose patent work focuses on machine learning methods for medical imaging analysis. His invention addresses how models can learn to identify anatomical objects when training datasets are only partially labeled.

Latest Patents

Liu is associated with 1 patent: “Learning from Partially Labeled Datasets for Medical Imaging Analysis.” The patent describes systems and methods that receive medical images, use a machine-learning model to perform an analysis task, and output the results. To train the model, the method uses training images and available ground-truth labels, then generates pseudo labels with the trained model and uses both the original labels and pseudo labels to fine-tune it. This approach is designed to support medical image analysis when complete annotations are unavailable.

Career Highlights

Liu works at Siemens Healthineers AG, a company active in medical technology and healthcare innovation. His patent reflects an application of machine learning to medical imaging, particularly the use of pseudo labels to make better use of partially labeled training data.

Collaborations

Liu’s listed coworker is Sasa Grbic. The patent information provided does not specify whether Grbic collaborated on this particular invention.

Conclusion

Han Liu’s patent presents a machine-learning approach for medical imaging analysis that combines ground-truth labels with generated pseudo labels, highlighting his contribution to innovation in healthcare technology.