Data/Software/Code

Archived Processed Results:
We are happy to announce that we will release the processed results on datasets OASIS3, ADNI, HCP, dHCP, BCP, ABIDE… The processed results (including skull stripping, tissue segmentation and parcellation) were generated by the volume-based analysis pipeline (integrated in iBEAT V2.0 Cloud, http://www.ibeat.cloud), which was designed in the Developing Brain Computing Lab (https://liwang.web.unc.edu) led by Dr. Li Wang (li_wang@med.unc.edu).

    • Before requesting the processed results, please carefully read and sign the Date Use Agreement of the Processed Results of ADNI. Once you have successfully submitted your request and uploaded the signed Date Use Agreement, you will be able to download the data. >>> Request for Archived Processed ADNI (9,500+ scans) 
    • Before requesting the processed results, please carefully read and sign the Date Use Agreement of the Processed Results of BCP. Once you have successfully submitted your request and uploaded the signed Date Use Agreement, you will be able to download the data. >>> Request for Archived Processed BCP (700+ scans) (in final packaging)

Online infant processing pipeline: iBEAT V2.0 Cloud

  • iBEAT V2.0 Cloud, http://www.ibeat.cloud/. iBEAT V2.0 Cloud is a toolbox for processing infant brain MR images, using multimodality (including T1w and T2w) or single-modality. Main functions of the software (step by step) include image preprocessing, brain extraction, tissue segmentation and brain labeling. The software is developed by the IDEA group at the University of North Carolina at Chapel Hill. iBEAT was first developed in 2012, now re-developed with more advanced techniques. So far, we have successfully processed 18,000+ infant brain images from multiple sites with various protocols and scanners (Table 1).

Table 1. Successfully processed 18,000+ infant brain images from multiple sites with various protocols and scanners.

Codes/Softwares:

  • Volume-based Analysis of 6-month-old Infant Brain MRI for Autism BioMarker Identification and Early Diagnosis [PDF] [Code: Caffe prototxt] [Software]
  • LINKS: Learning-based multi-source IntegratioN frameworK for Segmentation of infant brain images. This novel method employs the random forest and auto-context model. [Journal version] [PPT] [Email request for the code]
  • Neonatal Brain MR Image Segmentation using Sparse Representation and Patch-Driven Level Sets, Neuroimage, 84, 141-158, 2014. [PDF] [PPT][Matlab code]
  • Longitudinally guided level sets for consistent tissue segmentation of neonates, Human Brain Mapping, 34(4), 956-972, 2013.[PDF][BibTex][Infant processing software].
  • Automatic Segmentation of Neonatal Images Using Convex Optimization and Coupled Level Sets, NeuroImage, 58:805-817, 2011. [PDF][Infant processing software]
  • Medical Image Segmentation with Local Gaussian Distribution (LGD) Fitting Energy [PDF][Matlab Source Code]

Datasets:

  • One zip file with training/testing images and manual labels is available for download. The zip file contains T1- and T2-weighted MR images of 39 infant subjects from multiple sites (the challenge is always open):

  • One zip file with training images and manual labels is available for download. The zip file contains T1- and T2-weighted MR images of 10 infant subjects (the challenge is always open):