Underwater Object Detection and Recognition Using Yolo (You Only Look Once) Algorithm

Abstract

Similar to other projects in the subject of computer vision, object detection is a challenging task. Concurrent training
takes place in many image areas for both detection and classification. Convolutional models share computation
demands by distributing work among various sites. In contrast to localization tasks, object detection necessitates taking
into consideration a background class in the absence of an object. Finding and classifying areas of images both add to
the difficulty of object identification. Knowing how to divide images and locate objects is essential for successful
object detection. Knowing the position of an object helps determine its shape, and knowing the shape of an object
helps determine its location. For example, distinctive features, like a person's face and clothing, may be components
of the same entity. However, without initially identifying the thing, it is still challenging to understand its identity.
The objective is to create a very accurate and effective deep learning system that uses the YOLOv8 (You Only Look
Once) algorithm to identify a range of aquatic life organisms underwater.