《计算机应用》唯一官方网站 ›› 2024, Vol. 44 ›› Issue (4): 1114-1120.DOI: 10.11772/j.issn.1001-9081.2023081042

• 人工智能 • 上一篇    

基于注意力机制的鸟类识别算法

陈天华1, 朱家煊1(), 印杰2   

  1. 1.北京工商大学 人工智能学院,北京 100048
    2.江苏警官学院 计算机信息与网络安全系,南京 210031
  • 收稿日期:2023-08-08 修回日期:2023-12-04 发布日期:2023-12-18 出版日期:2024-04-10
  • 通讯作者: 朱家煊
  • 作者简介:陈天华(1966—),男,湖南长沙人,教授,硕士,主要研究方向:图像处理、模式识别、测控技术
    朱家煊(1997—),男,江苏南通人,硕士研究生,主要研究方向:模式识别、图像处理 408199640@qq.com
    印杰(1977—),男,江苏南京人,高级工程师,硕士,主要研究方向:机器学习、大数据、网络安全。
  • 基金资助:
    国家自然科学基金资助项目(62272203);北京市自然科学基金-北京市教育委员会科技计划重点项目联合项目(KZ202110011015)。

Bird recognition algorithm based on attention mechanism

Tianhua CHEN1, Jiaxuan ZHU1(), Jie YIN2   

  1. 1.School of Artificial Intelligence,Beijing Technology and Business University,Beijing 100048,China
    2.Department of Computer Information and Cybersecurity,Jiangsu Police Institute,Nanjing Jiangsu 210031,China
  • Received:2023-08-08 Revised:2023-12-04 Online:2023-12-18 Published:2024-04-10
  • Contact: Jiaxuan ZHU
  • About author:CHEN Tianhua, born in 1966, M. S., professor. His research interests include image processing, pattern recognition, measurement and control technology.
    ZHU Jiaxuan, born in 1997, M. S. candidate. His research interests include pattern recognition, image processing.
    YIN Jie, born in 1977, M. S., senior engineer. His research interests include machine learning, big data, cybersecurity.
  • Supported by:
    National Natural Science Foundation of China(62272203);Joint Project of Beijing Natural Science Foundation and Beijing Municipal Education Commission(KZ202110011015)

摘要:

针对现有细粒度鸟类目标识别算法准确率不高的问题,提出一种鸟类目标检测算法YOLOv5-Bird。首先,在YOLOv5主干网络中引入基于混合域的坐标注意力(CA)机制,增大有价值的通道权重,以区分目标特征和背景中的冗余特征;其次,在原始主干网络中采用双层路由注意力(BRA)模块替换原网络中的部分C3模块,过滤低相关度的键值对信息,获得高效的长距离依赖关系;最后,使用WIoU(Wise-Intersection over Union)损失函数,增强算法对目标的定位能力。实验结果表明,YOLOv5-Bird在自建数据集上取得了82.8%的精确率和77.0%的召回率,比YOLOv5算法分别提高4.3和7.6个百分点,也优于增加其他注意力机制的算法。验证了YOLOv5-Bird在鸟类目标检测场景中具有较好的性能。

关键词: 目标检测, 生物识别, 卷积神经网络, 注意力机制, 损失函数

Abstract:

Aiming at the low accuracy problem of existing algorithms for fine-grained target bird recognition tasks, a target detection algorithm for bird targets called YOLOv5-Bird, was proposed. Firstly, a mixed domain based Coordinate Attention (CA) mechanism was introduced in the backbone of YOLOv5 to increase the weights of valuable channels and distinguish the features of the target from the redundant features in the background. Secondly, Bi-level Routing Attention (BRA) modules were used to replace part C3 modules in the original backbone to filter the low correlated key-value pair information and obtain efficient long-distance dependencies. Finally, WIoU (Wise-Intersection over Union) function was used as loss function to enhance the localization ability of algorithm. Experimental results show that the detection precision of YOLOv5-Bird reaches 82.8%, and the recall reaches 77.0% on the self-constructed dataset, which are 4.3 and 7.6 percentage points higher than those of YOLOv5 algorithm. Compared with the algorithms adding other attention mechanisms, YOLOv5-Bird also has performance advantages.It is verified that YOLOv5-Bird has better performance in bird target detection scenarios.

Key words: target detection, biological recognition, Convolutional Neural Network (CNN), attention mechanism, loss function

中图分类号: