Dataset Card for Wilds-Beetles

This dataset includes 3,000+ hand taken images of approximately 300 beetles at the Wilds in July 2024. Aim is to succesfully build a model capable of idnetifying unique beetles based on spot patterns

Dataset Details

Dataset Structure

/dataset/ <img_id 1>.png <img_id 2>.png ... .png beetle_labels.csv

Data files

All images are named .png. They are 1024 x 1024. Images are mapped to beetles in beetle_labels.csv

beetle_labels.csv:

• fname: name for a given image. Ex IMG_0004.
• viewpoint: Portion of beetle visible to camera. [Top, back, head, left_side, right_side]
• 'beetle_id': The beetle which the image is taken of
• 'file_name': Full file name, which is fname + .JPEG i.e IMG_0004.JPEG
• species: Species of the specimen in the image. For this dataset, there is only one. [burying_beetle]
• file_path: Relative path to image from the root of the directory (<dataset>/<species>/<img_id>.jpeg); -->

Data Splits

[Work In Progress]

Curation Rationale

The wilds releases hundreds of beetles each year, some of which are ineviably lost track of or are hard to indentify due to manual tagging, which can be lost or damaged The ideal engame is to create a model or AI capable of identifying unique beetles based only on spot patterns using a single image, and having that model quickly identify a beetle given a new image once captured. This could eventually be done on edge (i.e on the same phone the image was taken on)

Data Collection and Processing

Data was mnaully collected over a period of several weeks. Beetles were indivudally handled by Michael and photos were taken at various viewpoints

Bias, Risks, and Limitations

There is no set standard for how many images of a givne beetle were taken, although I tried to capture 5+ covering several angles for each beetle. Some beetles were more cooperative than others, sometimes only top angle was succesfully captured without getting hands in image (i.e beetle would not sit still)

Licensing Information

CCO: This dataset is dedicated to the public domain for the benefit of scientific pursuits. We ask that you cite the dataset and journal paper using the below citations if you make use of it in your research.

This dataset (the compilation) has been marked as dedicated to the public domain by applying the CC0 Public Domain Waiver. However, images may be licensed under different terms (as noted above).

Citation

[More Information Needed]

BibTeX:

Acknowledgements

This work was supported by the Imageomics Institute, which is funded by the US National Science Foundation's Harnessing the Data Revolution (HDR) program under Award #2118240 (Imageomics: A New Frontier of Biological Information Powered by Knowledge-Guided Machine Learning). Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Glossary

More Information

Dataset Card Authors

@mdsilly

Dataset Card Contact

[More Information Needed--optional]