Awakening a taxonomist's third eye: exploring the utility of computer vision and deep learning in insect systematics

Abstract

We tested deep learning and computer vision methods for identifying all species of the genus Oxythyrea. Convolutional neural networks were able to distinguish all of them with better accuracy than human experts.

We tested a sex identification of mentioned beetles as well based only on dorsal view, which is uninformative for human expert in this task.

Keywords

• Hexapoda
• Oxythyrea
• computer vision
• deep learning
• insect systematics