README.md

metadata

library_name: mmsegmentation
pipeline_tag: image-segmentation
tags:
  - methane
  - gas-detection
  - optical-gas-imaging
  - segmentation
  - gasformer

Gasformer

Checkpoints for Gasformer: A Transformer-based Architecture for Segmenting Methane Emissions from Livestock in Optical Gas Imaging (arXiv 2404.10841).

• Code: https://github.com/toqitahamid/Gasformer

Folder	Dataset	Checkpoint	Config
CR/	Dairy Cow Rumen gas (binary)	CR_best_mIoU_iter_160000.pth	gasformer-160k_CR-512x512.py
MR/	Controlled Methane Release (multi-class)	MR_best_mIoU_iter_160000.pth	gasformer-160k_MR-512x512.py

Each checkpoint is the best-mIoU checkpoint from a 160k-iteration MMSegmentation run at 512x512.

Usage

Clone the code repository, then place the checkpoint next to its config:

git clone https://github.com/toqitahamid/Gasformer.git
hf download toqi/Gasformer --local-dir weights
python tools/test.py weights/CR/gasformer-160k_CR-512x512.py weights/CR/CR_best_mIoU_iter_160000.pth

Citation

@InProceedings{Sarker_2024_CVPR,
    author    = {Sarker, Toqi Tahamid and Embaby, Mohamed G and Ahmed, Khaled R and Abughazaleh, Amer},
    title     = {Gasformer: A Transformer-based Architecture for Segmenting Methane Emissions from Livestock in Optical Gas Imaging},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
    month     = {June},
    year      = {2024},
    pages     = {5489-5497}
}