Anomalous behviour detection for under water fish using AI techniques

In this work, detection of fish in a water and behavior like whether it’s in normal/abnormal posture is performed. Anomalous events detection in real-world video scenes is a challenging problem owing to the complexity of anomaly and the untidy backgrounds and objects in the scenes. 

Although there are already many studies on dealing with this problem using deep neural networks, very little literature aims for real-time detection of the anomalous behavior of fish. This paper presents an underwater fish anomalous behavior detection method by using deep learning techniques like YOLO object detector and convolutional neural network (CNN). 

YOLO is used to detect and track the fish under water where CNN is used to classify the posture of fish. This work will be useful for researchers and the aquaculture industry.

Keywords: Anomalous behavior analysis, Deep learning, Object detection, Tracking, Convolution neural network.