计算机与数字工程2025，Vol.53Issue(1)：250-256,7.DOI:10.3969/j.issn.1672-9722.2025.01.045
基于图像显著性迁移学习的水面泛目标检测算法
Surface Generic Object Detection Algorithm Based on Image Saliency Transfer Learning
摘要
Abstract
In the current water surface target detection task,researchers pay more attention to the characteristics of the target itself,and use the target data with similar characteristics for training,resulting in limitations of the algorithm in different target de-tection.Aiming at the problem of multi-target detection in water environment,this paper proposes a water surface generic target de-tection algorithm based on image transfer learning.The algorithm is based on the BASNet network model,and the final results are obtained by the up-down sampling operation and multi-level feature fusion of the input image through the prediction module and correction module of the model.The experimental results show that the algorithm has good detection effect on multiple targets on wa-ter surface,and the algorithm training model has migration learning ability.关键词
显著性目标检测/水面目标检测/特征提取/多级特征融合/迁移学习Key words
salient object detection/surface object detection/feature extraction/multi-level feature fusion/transfer learn-ing分类
信息技术与安全科学引用本文复制引用
董国霖,陈黎,陈姚节..基于图像显著性迁移学习的水面泛目标检测算法[J].计算机与数字工程,2025,53(1):250-256,7.基金项目
国家自然科学基金项目(编号:61773297)资助. （编号:61773297）
