Artificial Intelligence for Liver Ultrasound Analysis

Our ultrasound navigation platform can identify organs, veins, and lesions in abdominal ultrasound images of the liver. It can also detect and classify the severity of non-alcoholic fatty liver disease (NAFLD).

Our synthesis application generates liver ultrasound images using latent diffusion models: text-to-image, class-to-image, and semantic masks.

object detection 1

Empowering and promoting liver health with AI precision


AI for Classification and Navigation of Liver Ultrasound in Patients with Hepatocellular Carcinoma

Research presented at EASL Liver Cancer Summit 2023

  • Multi-model approach using object detection and semantic segmentation to delineate liver and kidney
  • Able to classify severity of steatosis using only images
  • Aid medical practitioners in liver protocol navigation and selection of region of interest

Tech Demos

Ultrasound Navigation Platform

object detection 2

Liver Image Synthesis

diffusion model

Generate realistic, synthetic
liver ultrasound images

Our app uses diffusion models to create ultrasound images that mimic the structural anatomy of the liver.


We are researchers from the UC Berkeley Master of Information and Data Science program aiming to make an impact in the medical imaging space.