Xray Projectomic Reconstruction Extracting Segment with Skeleton Banner
Join

XPRESS: Xray Projectomic Reconstruction - Extracting Segmentation with Skeletons (ISBI 2023)

This challenge is part of ISBI 2023, taking place on April 18-21th, 2023.  The top teams will be invited for writing an overview paper after the challenge. We will keep the online evaluation open for new submissions after the workshop. More information about the challenge can be found in our paper.

Visualize the dataset in your browser!

[Training raw + skeleton GT + voxel GT]  [Validation raw + skeleton GT] [Test raw]

Task

In this task, we provide volumetric XNH images of cortical white matter axons from the mouse brain at 100 nm per voxel isotropic resolution. Additionally, we provide ground truth annotations for axon trajectories. Manual voxel-wise annotation of ground truth is a time-consuming bottleneck for training segmentation networks. On the other hand, skeleton-based ground truth is much faster to annotate, and sufficient to determine connectivity. Therefore, we encourage participants to develop methods to leverage skeleton-based training. To this end, we provide two types of training (validation) sets: a small volume of voxel-wise annotations and a larger volume with skeleton-based annotations. The participants will have the flexibility to use either or both of the provided annotations to train their models, and are challenged to submit an accurate voxel-wise prediction on the test volume. Entries will be evaluated on how accurately the submitted segmentations agree with the ground-truth skeleton annotations.

Dataset [code]

Evaluation [code]

(a) XNH image data. (b) Example axon segmentation from the baseline model. (c) Examples of split error and  merge error.

Starter Code

  • Deep learning baseline segmentation model: [code]
  • Evaluation and submission utility functions: [code]
  • Google colab tutorial: [link]

Timeline

  • February 6, 2023. Release of the training dataset, validation dataset, and code for evaluation.
  • February 6, 2023. Release of test dataset and baseline models.
  • February 6, 2023 - April 3, 2023 (11:59 pm EST). Submission of test results  (ddl extended).
  • April 4, 2023. Final leaderboard release.
  • April 18, 2023. On-site workshop presentations at ISBI.
@article{nguyen2023xpress,
title={The XPRESS Challenge: Xray Projectomic Reconstruction--Extracting Segmentation with Skeletons}, author={Nguyen, Tri and Narwani, Mukul and Larson, Mark and Li, Yicong and Xie, Shuhan and Pfister, Hanspeter and Wei, Donglai and Shavit, Nir and Mi, Lu and Pacureanu, Alexandra and others}, journal={arXiv preprint arXiv:2302.03819}, year={2023} }

Organizers

  • Tri Nguyen, Department of Neurobiology, Harvard Medical School
  • Mukul Narwani, Department of Neurobiology, Harvard Medical School 
  • Mark Larson, College of Science, Northeastern University
  • Yicong Li, John A. Paulson School of Engineering and Applied Sciences, Harvard University 
  • Shuhan Xie, College of Science, Northeastern University
  • Hanspeter Pfister, John A. Paulson School of Engineering and Applied Sciences, Harvard University 
  • Donglai Wei, Department of Computer Science, Boston College
  • Nir Shavit, Computer Science & Artificial Intelligence Laboratory, MIT 
  • Lu Mi, Computer Science & Artificial Intelligence Laboratory, MIT 
  • Alexandra Pacureanu, ESRF, The European Synchrotron    
  • Wei-Chung Lee, Department of Neurobiology, Harvard Medical School & F.M. Kirby Neurobiology Center, Boston Children's Hospital, Harvard Medical School 
  • Aaron Kuan, Department of Neurobiology, Harvard Medical School

Acknowledgment

This project received funding from the NIH (EB032217) to ATK and (MH128949) to WCAL and AP, and the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement n°852455) to AP. We acknowledge the ESRF for granting beamtime for proposal LS2892.