Introduction

Foundation models are becoming increasingly important in medical image analysis as they offer advanced AI neural networks trained on extensive and diverse dataset. These models can be utilized for a wide range of tasks, including classification, detection, and segmentation.

Visual Foundation Models for Medical Image Analysis

Visual foundation models are emerging to address various problems in medical image analysis, opening up a world of possibilities for applications that enhance segmentation tasks and enable more accurate measurements. The MONAI Model Zoo provides access to a variety of pre-trained models for different medical imaging tasks such as whole-body CT segmentation and whole-brain MRI segmentation.

Whole-body CT segmentation

The MONAI team has developed models capable of segmenting all 104 anatomical structures from a single whole-body CT scan model. This includes 27 organs, 59 bones, 10 muscles, and 8 vessels. The available foundation model offers fast inference times and can be accessed in the MONAI Model Zoo.

Whole-brain MRI segmentation

The process of whole-brain segmentation is critical in medical image analysis and offers a non-invasive method to measure brain regions. The foundation model developed by the MONAI team and Vanderbilt University can simultaneously segment all 133 brain structures from MRI 3D segmentation.

How to access medical imaging foundation models

Fine-tuning pre-trained models is crucial to the future of medical image analysis. The MONAI Model Zoo offers access to a diverse collection of pre-trained models for various medical imaging tasks that can be fine-tuned or used as a starting point for specific applications. The use of foundation models in medical image analysis has great potential to improve diagnostic accuracy and enhance patient care.