An Underwater Species Detection with YOLOv9 Combining Self Attention Mechanism

Underwater species detection is a challenging task due to issues like scattering and absorption, which significantly reduce image clarity. This paper utilizes the YOLOv9 model, renowned for its real-time object detection capabilities, also incorporating a self- attention mechanism to enhance feature recognition. YOLOv9 introduces two key innovations: PGI (Programmable Gradient Information) and GELAN (Generalized Efficient Layer Aggregation Network). PGI enhances the model's ability by fine-tuning gradient flows, which preserves critical information during the learning process. GELAN enhances the detection process accuracy by dynamically merging layers to optimize the model's design. By integrating a self- attention mechanism, YOLOv9 further strengthens its performance by enabling the model to selectively focus on key areas of an image. Together, these features make YOLOv9 well-suited to overcoming the unique challenges of underwater environments, ensuring high detection accuracy.