AI Mosquito Alert Challenge Dataset 2023

The AI Mosquito Alert 2023 Challenge Dataset is a curated dataset used in the AI Mosquito Alert 2023 Challenge (https://www.aicrowd.com/challenges/mosquitoalert-challenge-2023), aimed at improving mosquito species identification through AI and deep learning models.

This dataset is distributed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC BY-NC-SA 4.0): https://creativecommons.org/licenses/by-nc-sa/4.0/.

The dataset was created through the efforts of the Mosquito Alert team, collaborators and thousands of citizen scientists. Please credit the Mosquito Alert Community (www.mosquitoalert.com) if you use this dataset (e.g., 'Mosquito Alert dataset, downloaded from [link], CC BY-NC-SA 4.0').

The intellectual property (IP) rights of this dataset belong to the Mosquito Alert team. 

The license is included in the file license.txt within the dataset zip file, along with the images, labels and dataset description.

The dataset consists of 10357 labeled images (approximately 9.8 GB in total). Images are accompanied by a designated CSV file called: annotations.csv. The CSV files include bounding box coordinates in the format: top left and bottom right notation ("bbx_xtl", "bbx_ytl", "bbx_xbr", "bbx_ybr").

The dataset consists of six distinct classes, including species and genus levels as well as a species complex. 

A summary of the mosquito classes, their descriptions, and corresponding class names used in the dataset:

• Aedes aegypti (species level) - class name: "aegypti"

• Aedes albopictus (species level) - class name: "albopictus"

• Anopheles (genus level) - class name: "anopheles"

• Culex (genus level) - class name: "culex" (species classification is challenging, so it is given at the genus level)

• Culiseta (genus level) - class name: "culiseta"

• Aedes japonicus/Aedes koreicus (species complex - difficult to differentiate between the two species) - class name: "japonicus-koreicus"

Label file:

The dataset includes a single CSV file: annotations.csv, which contains all the annotations for the images. Each row in the file provides the following information:

• img_fName: image file name

• img_w: image width

• img_h: image height

• bbx_xtl: bounding box top-left x-coordinate

• bbx_ytl: bounding box top-left y-coordinate

• bbx_xbr: bounding box bottom-right x-coordinate

• bbx_ybr: bounding box bottom-right y-coordinate

• class_label: class label (e.g., 'albopictus').

Additional notes:

• a broader description of the dataset and classes will be provided in the https://www.aicrowd.com/challenges/mosquitoalert-challenge-2023#dataset and https://www.youtube.com/watch?v=qSWJZUY-5DM challenge video

• exif information has been removed from the images for privacy protection

• most images contain a single mosquito with its corresponding bounding box and class label. However, in rare cases with multiple mosquitoes, only one mosquito is assigned a bounding box and label for consistency and compatibility.