This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.

Data Science Challenge 2025

Full Dataset Release: MOVi-MC-AC

Welcome to the 2025 Lawrence Livermore National Laboratory Data Science Challenge!

Be sure to git-lfs install before cloning the repo, otherwise the large files (managed by LFS) won't be cloned properly!

git-lfs install
git clone https://huggingface.co/datasets/Amar-S/LLNL_DSC_2025

See MOVi-MC-AC/DSC_2025.ipynb to get started!

This notebook serves as the technical introduction to LLNL's MOVi-MC-AC Dataset (last bullet above), covering:

(1) Introduction to Image Processing / Computer Vision
(2) Example Baseline Experiment
(3) DSC Challenge & Tasking
- Task 1.1: (Image-based) Modal Mask -> Amodal Mask
- Task 1.2: (Image-based) Modal Content (RGB) -> Amodal Content (RGB)
- Task 2.1: (Video-based) Modal Mask -> Amodal Mask
- Task 2.2: (Video-based) Modal Content (RGB) -> Amodal Content (RGB)
- Transfer Test: Apply Models on DSC Target Dataset:
  - Gather Modal Masks from some SotA method (SAM2)
  - Predict Amodal Masks, using Modal Masks
- Bonus Task 3: Create Modal Masks with SAM2
- Bonus Task 4: Re-ID of Objects