NiChart is a novel AI-powered neuroimaging platform with tools for computing a dimensional chart from multi-modal MRI data. NiChart provides end-to-end pipelines from raw DICOM data to advanced AI biomarkers, allowing to map a subject’s MRI images into personalized measurements, along with reference distributions for comparison to a broader population.
This repo contains the NiChart application front-end, which ties together all individual tools in the NiChart ecosystem and provides an easy-to-use interface for processing your data.
The development of NiChart is guided by several core principles:
- Enabling near real-time image processing and analysis through advanced methods.
- Facilitating the continuous integration of cutting-edge methods for extracting novel AI biomarkers from neuroimaging data.
- Ensuring robust and reliable results through extensive data training and validation on large and diverse training datasets.
- Providing user-friendly tools for visualization and reporting.
- Developing a deployment strategy that enables easy access for users with varying technical expertise and hardware resources.
We provide both a locally installable desktop application and a cloud-based application.
The NiChart cloud application, hosted via Amazon Web Services (AWS), deploys scalable infrastructure which hosts the NiChart tools as a standard web application accessible via the user’s web browser. No payment or installation is needed to use the tool.
However, as a web application, NiChart Cloud requires you to upload your data to the private cloud-based NiChart server for us to process it. We do not access or use your data for any other purpose than to run your requested processing and/or provide support to you as a user, and we regularly automatically delete user data after inactivity. However, we recognize that data privacy agreements and related concerns may nevertheless restrict use of the cloud application. If that applies to you, we suggest that you install the desktop application. We provide detailed installation instructions on the Installation page.
Want to switch between versions? The cloud and desktop applications are unified at the code level through the use of the Python library Streamlit. Consequently, the user experience is nearly identical between the cloud and desktop applications.
If you're looking for the individual NiChart structural tools, please see their individual repos:
NiChart_DLMUSE [GitHub] [Docker Hub] - Fast brain segmentation via deep learning
NiChart_DLWMLS [GitHub] [Docker Hub] - Fast white matter lesion segmentation via deep learning
SPARE scores [GitHub] [Docker Hub] - ML-based prediction/scoring for variables of clinical interest
CCL-NMF prediction [GitHub] [Docker Hub] - Lightweight estimation of CCL-NMF loading coefficients
ComBatFam Harmonization [GitHub] [Docker Hub] - Dataset harmonization tools
SurrealGAN / PredCRD [GitHub] [Docker Hub] - Predict continuous representation of disease along 5 principal dimensions
© 2025 CBICA. All Rights Reserved.
