This study presents an AI-based system for evaluating bee health, using a combination of visual and audio signals. The system’s core is an Attention-based Multimodal Neural Network (AMNN), which focuses on key features from each type of signal for accurate bee health assessment. The system outperformed existing models in accuracy and efficiency, and it proved that audio signals are more reliable than images for assessing bee health. This approach offers a more efficient, non-invasive solution for early detection of bee diseases and preservation of bee colonies.
Publication date: 18 Jan 2024
Project Page: arXiv:2401.09988v1
Paper: https://arxiv.org/pdf/2401.09988
