[2O4-GS-13] AI application: Agriculture, forestry and fisheries

[2O4-GS-13-05] Multiple Honeybee Detection from Hive Image Based on Deep Autoencoder and Transfer Learning

〇Shinya Takahashi¹, Yujie Li¹, Naoyuki Tsuruta¹, Hiroyuki Ai¹ (1. Fukuoka University)

Keywords:Deep Representation Learning, Transfer Learning, Object Tracking, honeybee, Computational Ethology

Recently, a new research field "Computational Ethology" is attracting much attention from not only biologists but also computer scientists because advances in computer technology enabled to automate the measurement and the analysis of animal behavior. Especially, computer-aided analyzing of communications performed by honeybee workers in their hive is one of the most important and interesting issue in ethological research area to reveal a mechanism of honeybee's language. These analyses have been usually conducted by manually extracting honeybee's walking trajectories from long time video data. For a systematic and theoretical analysis of honeybee's communications, we have developed an automatic tracking algorithm of multiple honeybees using image processing. However, the detection accuracy of the honeybee abdomen region is about 80% and further high precision is desired. So, in this research, we attempt to detect the honeybee region based on a deep learning approach using deep autoencoder and transfer learning. In the proposed method, high accurately classification of the honeybee images is achieved using the feature vectors obtained by the deep autoencoder instead of the Haar-like features in the previous work without preparing enormous labeled data in advance. In this paper, we show the experimental results and confirm the proposed model's capabilities.