Author：Fisheries Research Institute,COA

Marine ornamental fish with beautiful shapes and gorgeous colors are extremely popular in the aquarium market and play a vital role in the ecosystem. The behavior image recognition system, as developed in this study, identifies the breeding fish of the blackfinned clownfish and provides an innovative way to use image processing and recognition technologies to facilitate obtaining accurate and reliable tracking data in a controllable experimental environment. Marine ornamental fish usually show territorial behaviors during breeding, which causes injuries to precious fish species. In conventional practice, to prevent injuries and even the death of fish, a lot of manpower is required to observe the breeding behaviors in real time. This study uses the breeding fish image recognition system to track and record the reproductive interactions and predation behaviors of sea anemones in a simulated experiment. Incidents are analyzed and calculated on this basis to facilitate maintenance, management, and formulation of management strategies.