Detecting Fish Species Through Mask RCNN

Detection of underwater fish species has become more important through marine science research. Automatic detection of fish species would facilitate the development of marine science. There are several different methods that can be employed to determine fish species. Each of the available methods requires a substantial amount of training data to detect precisely. To overcome this, in this article, we report on the findings of an investigation that was carried out utilizing an deep learning algorithm known as Mask Region based Convolution Neural Network (RCNN). The investigation was conducted using a dataset containing images of various types and directions of fish found in different parts of the world. Mask RCNN is able to produce the highest quality segmentation mask possible for every occurrence and different fish species with a high degree of accuracy. The mask RCNN is able to recognize and categorize fish species with an overall accuracy of almost 97%.