基于改进YOLOv8s的肺结节检测算法
(西南交通大学 计算机与人工智能学院,四川 成都 611756)
摘要
肺癌已成为世界范围内死亡率较高的常见癌症之一,肺结节是肺癌的早期表现。针对CT影像中肺结节检测困难的问题,提出一种基于改进YOLOv8s的肺结节检测算法。主干网络采用改进的YOLOv8s,使用空间深度转换卷积替换传统的步长卷积和池化层,避免步长卷积和池化层在处理低分辨率图像或小物体时导致细粒度信息的丢失。加入坐标注意力模块,考虑了肺结节影像的通道间关系和位置信息,使模型能更准确地定位并识别目标区域。使用Focal Loss替换交叉熵损失函数,以解决数据集样本标签数量较少的问题。采用自适应激活函数,不仅能提升网络的稳定性,还能提高网络精度。使用LUNA16数据集验证算法性能,改进后的肺结节检测算法检测精度达到了77.8%,较Faster R-CNN和YOLOv8s检测算法分别提升了3.7%和8.6%。
关键词
肺结节检测;YOLOv8s;无步长卷积;自适应激活函数;空间注意力机制
参考文献
- [1] SHARMA P,MEHTA M,DHANJAL D S,et al. Emerging Trends in the Novel Drug Delivery Approaches for the Treatment of Lung Cancer [J/OL].Chemico-Biological Interactions,2019,309:108720[2024-10-03]. https://doi.org/10.1016/j.cbi.2019.06.033.
- [2] SCHABATH M B,COTE M L. Cancer Progress and Priorities: Lung Cancer [J]. Cancer Epidemiology, Biomarkers & Prevention,2019,28(10):1563-1579.
- [3] GIRSHICK R,DONAHUE J,DARRELL T,et al. Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation [C]//2014 IEEE Conference on Computer Vision and Pattern Recognition.Columbus:IEEE,2014:580-587.
- [4] GIRSHICK R. Fast R-CNN [C]//2015 IEEE International Conference on Computer Vision (ICCV).Santiago:IEEE,2015:1440-448.
- [5] LIU W,ANGUELOV D,ERHAN D,et al. SSD: Single Shot MultiBox Detector [C]//Computer Vision - ECCV 2016.Amsterdam:Springer,2016:21-37.
- [6] TERVEN J,CÓRDOVA-ESPARZA D M,ROMERO-GONZÁLEZ J A. A Comprehensive Review of YOLO Architectures in Computer Vision: From YOLOv1 to YOLOv8 and YOLO-NAS [J].Machine Learning & Knowledge Extraction,2023,5(4):1680-1716.
- [7] ZHU X Z,SU W J,LU L W,et al. Deformable DETR: Deformable Transformers for End-to-End Object Detection [J/OL].arXiv:2010.04159 [cs.CV].[2024-09-22].https://doi.org/10.48550/arXiv.2010.04159.
- [8] VARGHESE R,SAMBATH M.YOLOv8: A Novel Object Detection Algorithm with Enhanced Performance and Robustness [C]//2024 International Conference on Advances in Data Engineering and Intelligent Computing Systems (ADICS).Chennai:IEEE,2024:1-6.
- [9] SUNKARA R,LUO T. No More Strided Convolutions or Pooling: A New CNN Building Block for Low-Resolution Images and Small Objects [C]//Machine Learning and Knowledge Discovery in Databases.Grenoble:Springer,2022:443-459.
- [10] HOU Q B,ZHOU D Q,FENG J S. Coordinate Attention for Efficient Mobile Network Design [C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Nashville:IEEE,2021:13713-13722.
- [11] ROSS T Y,DOLLÁR G. Focal Loss for Dense Object Detection [C]//IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,4(2):2980-2988.
- [12] MA N N,ZHANG X Y,LIU M,et al. Activate or Not: Learning Customized Activation [C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Nashville:2021:8032-8042.
作者简介
郭相均(1997—),男,汉族,重庆开州人,硕士在读,研究方向:计算机视觉;通信作者:蒋朝根(1960—),男,汉族,浙江金华人,教授,硕士,研究方向:嵌入式系统及其应用、无线传感器网路。