EU Moths Dataset

This folder contains the EU moths dataset prepared for the evaluation of fine-grained recognition approaches. If you use this dataset, please cite the corresponding paper:

Dimitri Korsch and Paul Bodesheim and Joachim Denzler. Deep Learning Pipeline for Automated Visual Moth Monitoring: Insect Localization and Species Classification Fine-grained Recognition Datasets for Biodiversity Analysis. INFORMATIK 2021, Computer Science for Biodiversity Workshop (CS4Biodiversity). Pages 443-460. 2021. DOI: 10.18420/informatik2021-036

For more details visit the website of the dataset: https://inf-cv.uni-jena.de/eu_moths_dataset/

Directory Information

• images/ This directory contains subdirectories named after the corresponding moth species and each subdirectory contains the sample images of that species

• images.txt The filenames of the images relative to the images directory Each line contains a separate image.

• labels.txt The class IDs for each image in images.txt in the range 0 to 199. Each line in labels.txt corresponds to the image in the same line in images.txt

• tr_ID.txt Proposed train-test-splits for each image in images.txt. Each line in tr_ID.txt corresponds to the image in the same line in images.txt. Values are integers ranging from 0 to 3 to and might indicate different test folds. For example, a value of 0 means, this image is used for testing, and a value different from 0 (1 to 3) means this image is used for training.

• class_names.txt List of class names for the 200 moth species. They correspond to the subdirectory names in images/

• subset1_class_names.txt and subset1_class_names.txt Proposed subsets of species as in the paper (Korsch et al., 2021)

• eu_moths_downloads.csv List of download links for retrieving additional training images for indicated species. Each line corresponds to a species name and an URL for an image separated by a comma