Automatic Analysis of Fish Behaviors and Abnormality Detection

Current state-of-the-art fish monitoring systems are lack of intelligent in interpreting fishes behaviors automatically. To tackle these problems, we propose a vision-based method that automatically analyze behaviors of a group of fishes in an aquarium and detect abnormality precisely. Here we consider the problem in two steps. First, we propose a new incremental spectral clustering method to extract frequently occurred key swimming patterns of fishes. Then, we present video sequences of fishes into a trajectory through the space of these key patterns. Studying these trajectories provides a new tool to analyze fishes behaviors. Comparisons of fishes behaviors in clean water and water in the presence of chemicals provides a new tool to detect any abnormality. Experimental results illustrate that the precision value of our proposed method is above 90%.