Overview#
Description#
NVIDIA MAISI (Medical AI for Synthetic Imaging) is a state-of-the-art three-dimensional (3D) Latent Diffusion Model designed for generating high-quality synthetic CT images with or without anatomical annotations. This AI model excels in data augmentation and creating realistic medical imaging data to supplement limited datasets due to privacy concerns or rare conditions. It can also significantly enhance the performance of other medical imaging AI models by generating diverse and realistic training data.
In the latest Release 1.0.1, the underlying DDPM noise scheduler is replaced with Rectified Flow [1] for both training and inference. It accelerated latent diffusion model inference by 33x. The total speed of image generation can be 10x faster compared with the previous version 1.0.0.
MAISI offers several key features:
Generates high-resolution 3D CT images up to 512 × 512 × 768 voxels
Supports variable voxel sizes ranging from 0.5mm to 5.0mm
Capable of annotating up to 127 anatomical classes, including organs and tumors
Allows controllable anatomy size for 10 specific classes
Produces paired segmentation masks
(Release 1.0.1) Has better support for head region and small output volumes than Version 1.0.0.
(Release 1.0.1) Accelerated MAISI with Rectified Flow [1], can be 10x faster than Version 1.0.0.
By providing these capabilities, MAISI is a valuable tool for researchers advancing AI applications in healthcare. However, it is important to note that this model is intended for research purposes only and not for clinical usage.
MAISI Workflow#
The inference workflow of MAISI is depicted in the figure below. It first generates latent features from random noise by applying multiple denoising steps using the trained diffusion model. Then it decodes the denoised latent features into images using the trained autoencoder.
Advantage of MAISI NIM#
The NVIDIA NIM for MAISI simplifies the deployment of the MAISI model, offering several benefits for system administrators and developers:
Increased Productivity: The MAISI NIM allows developers to build generative AI applications quickly, in minutes rather than weeks, by providing a standardized way to add AI capabilities to their applications.
Simplified Deployment: The MAISI NIM provides containers that can be easily deployed on various platforms, including clouds, data centers, or workstations, making it convenient for developers to test and deploy their applications.
Scalability: The containerized approach ensures that the deployment can scale according to the needs of the application, whether it’s for a small research project or a large-scale deployment.
Consistency: Using NIMs ensures that the environment is consistent across different deployment platforms, reducing the chances of environment-related issues.
By leveraging the MAISI NIM, developers can focus more on innovation and less on the complexities of deployment, ensuring a smoother and more efficient workflow.
You are responsible for ensuring that your use of NVIDIA AI Foundation Models complies with all applicable laws.
Terms of Use#
By using this model, you are agreeing to the terms and conditions of the license.