EPBC-YOLOv8: An efficient and accurate improved YOLOv8 underwater detector based on an attention mechanism

TL;DR

This paper introduces EPBC-YOLOv8, an improved underwater object detection model that integrates attention mechanisms and advanced feature fusion techniques, achieving higher accuracy than the original YOLOv8 on underwater datasets.

Contribution

The study presents novel modifications to YOLOv8, including attention integration and enhanced feature fusion, specifically tailored for underwater target detection.

Findings

Achieved mAP of 76.7% on URPC2019 and 79.0% on URPC2020 datasets.

Outperformed original YOLOv8 by 2.3% and 0.7% in mAP scores.

Demonstrated improved detection accuracy for marine organisms.

Abstract

In this study, we enhance underwater target detection by integrating channel and spatial attention into YOLOv8's backbone, applying Pointwise Convolution in FasterNeXt for the FasterPW model, and leveraging Weighted Concat in a BiFPN-inspired WFPN structure for improved cross-scale connections and robustness. Utilizing CARAFE for refined feature reassembly, our framework addresses underwater image degradation, achieving mAP at 0.5 scores of 76.7 percent and 79.0 percent on URPC2019 and URPC2020 datasets, respectively. These scores are 2.3 percent and 0.7 percent higher than the original YOLOv8, showcasing enhanced precision in detecting marine organisms.