Feasibility Study of Fish Disease Detection using Computer Vision and Deep Convolutional Neural Network (DCNN) Algorithm

Detection of diseased fish at an early stage is necessary to prevent the spread of the disease. However, detecting fish diseases still uses a manual process and requires a high level of expertise that can be prone to human error. The ability of automatic detection of these fish diseases is much needed to help and to prevent losses of economic in the aquaculture industry. Therefore, this paper aims to detect disease of fish using computer vision and deep convolutional neural network (DCNN) algorithm. One Thousand and Two Hundred fish samples images were selected is namely diseased fish and healthy fish, which is determined by expert of fish diseases according to the specific of characteristics of fish diseases. The fish images went through the DCNN classifier and successfully achieved a satisfying mean average precision (mAP) with 0.237. The result shows that the computer vision integrated with the DCNN algorithm can efficiently predict fish disease.