Rotifer detection and tracking framework using deep learning for automatic culture systems

Abstract

Although rotifers (Brachionus plicatilis sp. complex) are an important first feed source in marine fish aquaculture, their management is quite time-consuming because their populations and movements need to be monitored daily. This management is still performed manually, and automation is required. If we could make good use of the recent breakthroughs in deep learning, the automating a rotifer culture system could be realized. We propose a deep learning framework for detecting and tracking rotifers as a basis for such automation and carefully verify its accuracy. Experimental results showed that a mean average precision of 88.5 % was achieved for detection, and a higher-order tracking accuracy of 88.7 % was achieved for tracking, indicating the suitability of deep learning methods for predicting the state of rotifers. In addition, this research will contribute to the development of the field by releasing the trained model and code for visualizing the tracking results, as well as an annotated dataset with over 30,000 instances.

Keywords

• Rotifer
• Automatic measurement
• Deep learning
• Object detection
• Multiple object tracking