LighterDet: A Lightweight Underwater Object Detection Method Based on YOLOv8

Underwater optical image object detection plays a crucial role in marine resource exploitation. However, due to limited computational resources in underwater computing devices, existing underwater object detection methods with numerous parameters are not suitable for such scenarios. To address this issue, we propose a new lightweight underwater object detection method based on YOLOv8, which is designed for resource-constrained scenarios. Specifically, we incorporate a lightweight backbone to reduce the detector's computational burden. Additionally, we propose a novel lightweight feature pyramid network that incorporates the depthwise separable convolutions. Extensive experiments are conducted to validate the effectiveness and efficiency of our method. The experimental results show that our method exceeds the state-of-the-art detector YOLOv8 by 1.8 AP points.