license: mit
task_categories:
- image-segmentation
tags:
- wsss
- boundary-refinement
- priors
PASCAL VOC 2012 NAMLab Priors for HGA
This dataset repository hosts the pre-computed hierarchical invariant region partition priors for weakly supervised semantic segmentation (WSSS), generated using the NAMLab framework.
These priors serve as the 2D structural guidance stream for the HGA (Hierarchical-Geometric Alignment) boundary internalization paradigm.
🔗 Main Codebase
The complete implementation, environment setup, and training guidelines are available in our official GitHub repository: 👉 Uncertainty-42/HGA
📂 Expected Directory Layout
Download the priors_voc.zip archive (approx. 4.2 GB) and extract its contents into your local PASCAL VOC dataset structure as shown below:
VOCdevkit/VOC2012/
├── JPEGImages/ # Original source images (.jpg)
├── SegmentationClassAug/ # Ground Truth masks (.png)
└── priors/
└── namlab_pt/ # <- Extract zip contents directly here (.pt files)
📥 Extraction Command
Navigate to your local priors/ directory and run:
unzip priors_voc.zip -d namlab_pt/
📝 Citation & Attribution
If you use these priors in your research, please cite both our work and the foundational NAMLab paper:
@article{zheng2021hierarchical,
title={Hierarchical Image Segmentation Based on Nonsymmetry and Anti-Packing Pattern Representation Model},
author={Zheng, Yunping and Yang, Bowen and Sarem, Mudar},
journal={IEEE Transactions on Image Processing},
volume={30},
pages={2408--2421},
year={2021},
publisher={IEEE}
}