TaxaNet: Harnessing a Hierarchical Loss Function for Insect Classification Using Deep Learning

Insects have the largest percentage of all living organisms worldwide, playing a pivotal role in maintaining essential ecosystem services such as pollination, pest control, nutrient cycling, and food provisioning. However, recent studies have reported alarming declines in insect populations globally, highlighting an urgent need for automated methods to estimate and quantify these populations, to better understand the reasons of their decline and to take proper measures. The ability to automatically estimate insect populations is crucial for shaping appropriate environmental policies. Advances in AI and computer vision techniques are revolutionizing the study of insects through non-invasive camera traps. However, the diversity of insects, close resemblances of many species, and multi-level taxa classification remain significant challenges in image-based insect monitoring. In this work, we propose TaxaNet a deep learning model for multi-level insect taxa classification, utilizing a pretrained EfficientNet as a feature extractor, followed by six classification blocks. Each block predicts one of the six taxonomic levels: Kingdom, Class, Order, Family, Genus, and Species. This hierarchical design and the loss function improves lower-level taxa predictions by leveraging the higher-level features available. A class-weighted hierarchical loss function, alongside the standard class-wise loss, allows the model to understand the relationships between taxonomic levels while maintaining classification accuracy. Trained on the Diopsis insect camera trap dataset containing 31,000 training images and 7,900 test images, the model achieved an average precision of 0.85 and a recall of 0.86 across five taxonomic levels. These results demonstrate the effectiveness of our approach in harnessing multi-level insect taxonomy to achieve multi-level insect classification.