Towards Explainable Visual Vessel Recognition Using Fine-Grained Classification and Image Retrieval

This repository shows how to download the Military MARVEL Dataset and sort it by class or superclass.

Here is the direct link to the paper: click me

The code can be excecuted using Python3.10, the dependencies can be found in requirements.txt and they can be installed using:

pip install -r requirements.txt

Downloading Military MARVEL dataset

In order to download the Military MARVEL dataset, run the following command:

python download.py

Sort images by class or superclass:

In order to sort the images of the Military MARVEL dataset, run the following command:

python sort_by.py

This will create folders for each unique class and puts the corresponding images in it. The structure can be used by ImageFolder class of PyTorch. Set SORT_BY = "class" if you want to create the fine-grained dataset with 137 classes and SORT_BY = "superclass" to create the less fine-grained dataset with 11 classes. The file structure of the GT file metadata.json looks as follows:

[{"image_id": "1337893", "superclass": "Auxiliaries", "class": "Schwedeneck class"}, {...}]

Cite us

If you use Military MARVEL in your research, please cite:

@InProceedings{Karus_2024_CVPR,
    author    = {Karus, Heiko and Schwenker, Friedhelm and Munz, Michael and Teutsch, Michael},
    title     = {Towards Explainable Visual Vessel Recognition Using Fine-Grained Classification and Image Retrieval},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
    month     = {June},
    year      = {2024},
    pages     = {82-92}
}