University of Michigan / Research notes
From cardiac images to structured measurements.
Automated cardiac chamber segmentation, ejection-fraction quantification, and a platform aimed at portable ultrasound interoperability.
The work, at a glance
I worked with Professor Joyce Wang on healthcare-driven AI algorithms. The research combined cardiac image analysis with a standardized cardiovascular disease (CVD) diagnosis platform.
Image analysis
Models for automated cardiac chamber segmentation and ejection-fraction quantification.
Platform integration
A standardized CVD diagnosis platform targeting portable ultrasound interoperability.
01 / Context
Connect image analysis to measurements.
The project focused on automated analysis of cardiac images. My model work covered segmentation of cardiac chambers and quantification of ejection fraction.
These are connected tasks: identifying structures in an image and producing measurements that can be represented in a consistent platform.
02 / Research components
Separate the model tasks from the platform.
I engineered models for chamber segmentation and ejection-fraction quantification. Alongside those models, I built a standardized CVD diagnosis platform.
From images to measurements
The image-analysis tasks produce structured measurements; the platform provides a common interface for using them.
03 / Platform direction
Design toward portable ultrasound.
Interoperability with portable ultrasound devices was a target of the platform. This extended the project’s scope beyond an isolated image-analysis model to the interface through which those capabilities could be used.
The work brought together model development and platform engineering within the Precision Imaging × AI Laboratory.