Data

Tools

Indexes

Applications

Experiments

Open Source Science

https://github.com/alleninstitute/mri_image_label_transfer

https://github.com/alleninstitute/mri_image_label_transfer

Science Score: 13.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ○
    .zenodo.json file
  • ○
    DOI references
  • ○
    Academic publication links
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (6.6%) to scientific vocabulary
Last synced: 9 months ago · JSON representation

Repository

Basic Info
  • Host: GitHub
  • Owner: AllenInstitute
  • Language: Python
  • Default Branch: main
  • Size: 0 Bytes
Statistics
  • Stars: 0
  • Watchers: 2
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created over 1 year ago · Last pushed over 1 year ago
Metadata Files
Readme

README.md

mriimagelabel_transfer

Visit the application website (deployed): https://alleninstitute-mri-image-label-transfer-my-app-6mlgh3.streamlit.app/

Watch the Demo Video: https://drive.google.com/file/d/1ZLDseIsR0dr1hZUTEyb7215lge7TRJAk/view?usp=share_link

Project Overview: This project was designed as a pilot for transferring brain region segmentation labels from sparsely labeled image volumes. Brain annotation is a time-intensive task that requires manual pixel-by-pixel labeling of structures defined by the human brain ontology. Because of time constraints, the most up-to-date brain annotation only covers ~10% of the brain volume. This project’s goal was to develop and deploy a model that transferred labeled sections of the Nissl human brain atlas to unlabeled neighbors. ​ We implemented a semi-supervised algorithm based on a kernel-density function described in Delalleau et al. (2005). We first trained a CNN on the voxel values of a Nissl-stained section and a paired segmentation volume. We then used the tensor embeddings to estimate the label in an unlabeled neighboring section. Our implementation was successful in labeling high contrast features such as the cortical grey matter. However, hyperparameter tuning is necessary for lower contrast regions before scaling to the whole brain volume. ​ ​ This project represents an important first step in label transfer for anatomical atlas images and will be useful for future atlas segmentation efforts in AIBS. ​

Owner

  • Name: Allen Institute
  • Login: AllenInstitute
  • Kind: organization
  • Location: Seattle, WA

Please visit http://alleninstitute.github.io/ for more information.

GitHub Events

Total
  • Member event: 1
  • Push event: 5
  • Create event: 2
Last Year
  • Member event: 1
  • Push event: 5
  • Create event: 2

Dependencies

requirements.txt pypi
  • numpy ==2.2.3
  • pandas ==2.2.3
  • pillow ==11.1.0
  • scikit-learn ==1.6.1
  • scipy ==1.15.2
  • streamlit ==1.42.2
  • torch ==2.6.0