Tensorflow dataset containing images of bees for multilabel classification
This dataset contains ~7.500 images of bees, captured at the entrance of a bee hive. Each image contains one bee, rotated to a vertical format, heads or tails up. This dataset was created for a masters thesis on detecting characteristics of bees and parasites like the varroa mite.
It is also part of the tensorflow dataset: https://www.tensorflow.org/datasets/catalog/bee_dataset
All images have multiple labels, such as:
- cooling_output - Indicating that the bee is currently cooling the hive. The bee is flapping its wings while keeping its position stationary, that way it transports fresh air into the hive.
- pollen_output - Indicates that the bee carries a pollen packet
- varroa_output - Indicates that the bee is infested with the varroa mite. The varroa mite is a small circular pest which is 1-2mm in diameter and of brown color. The mite will be carried into the hive, where it seriously damages the bee larva. If left untreated, the whole colony dies.
- wasps_output - Instead of a bee, a wasp was captured.