|
基于YOLOv2的微创手术机器人肺结核病灶图像识别研究
|
Abstract:
肺结核是全球公共卫生问题,每年数百万人受到其影响,其中包括成千上万的死亡病例。尽管存在有效的治疗方法,但及时诊断和早期干预仍然是防止病灶扩大的关键,其中外科手术治疗成为了一种有效手段。本文提出了一种基于YOLOv2算法的微创手术机器人肺结核病灶识别技术。该技术结合了微创手术机器人、内窥镜、K210视觉识别模块以及YOLOv2识别算法,以提高肺结核病灶实时监测的识别准确度和处理效率。本文构建了训练数据集,使用YOLOv2算法进行深度网络训练,并评估了性能指标,包括准确率、召回率、精确度和F1分数。实验结果表明,该技术具有高准确度和高效率,有望改进肺结核的检测与治疗手段。
Tuberculosis has been a major challenge in the global public health sector, affecting millions of people annually, including tens of thousands of fatalities. Despite the existence of effective treatment methods, timely diagnosis and early intervention remain crucial in preventing the spread of lesions. Surgical intervention has emerged as an effective means of treatment. This paper introduces a minimally invasive surgical robot-based tuberculosis lesion recognition technology, which is built upon the YOLOv2 algorithm. This technology combines robot system design, endoscopy, the K210 visual recognition module, and the YOLOv2 recognition algorithm to enhance the identification accuracy and the handling efficiency for real-time recognition. The paper constructed a diverse training dataset, conducted deep network training using the YOLOv2 algorithm, and evaluated performance metrics, including accuracy, recall, precision, and F1 score. Experimental results demonstrate that this technology offers high accuracy and efficiency, with the potential to enhance tuberculosis detection and treatment.
[1] | 窦学军, 龚昌帆, 闫东杰, 等. 手术联合化疗治疗耐多药肺结核结果分析[J]. 中国防痨杂志, 2007, 29(4): 336-338. |
[2] | 周胤江, 徐晓华. 肺结核手术治疗56例临床分析[J]. 内蒙古医学杂志, 2006, 38(4): 359. |
[3] | 刘龑龙, 刘进长, 唐莉, 等. 基于论文和专利的微创手术机器人领域研发现状及前沿热点分析[J]. 科学技术与工程, 2022, 22(1): 250-258. |
[4] | 付宜利, 潘博. 微创外科手术机器人技术研究进展[J]. 哈尔滨工业大学学报, 2019, 51(1): 1-15. |
[5] | 苏赫. 微创手术机器人图像系统设计方法与空间映射策略研究[D]: [硕士学位论文]. 天津: 天津大学, 2017. |
[6] | Park, B., Chi, H., Park, B., et al. (2023) Visual Modalities-Based Multimodal Fusion for Surgical Phase Recognition. Computers in Biology and Medicine, 166, Article ID: 107453. https://doi.org/10.1016/j.compbiomed.2023.107453 |
[7] | Lakhani, P. and Sundaram, B. (2017) Deep Learning at Chest Radiography: Automated Classification of Pulmonary Tuberculosis by Using Convolutional Neural Networks. Radiology, 284, 574-582.
https://doi.org/10.1148/radiol.2017162326 |
[8] | 单连峰, 郭芳. 基于深度学习卷积神经网络的肺结核CT辅助诊断模型在临床中的应用价值[J]. 中国新通信, 2022, 24(24): 126-128. |
[9] | 邹承明, 薛榕刚. 融合GIoU和Focal loss的YOLOv3目标检测算法[J]. 计算机工程与应用, 2020, 56(24): 214-222. |
[10] | 许德刚, 王露, 李凡. 深度学习的典型目标检测算法研究综述[J]. 计算机工程与应用, 2021, 57(8): 10-25. |
[11] | Zhou, H., Xiao, Y., Zheng, Z., et al. (2022) YOLOv2-Tiny Target Detection System Based on FPGA Platform. 2022 3rd International Conference on Big Data, Artificial Intelligence and Internet of Things Engineering (ICBAIE). Xi’an, 15-17 July 2022, 289-292. https://doi.org/10.1109/ICBAIE56435.2022.9985817 |
[12] | Ioffe, S. and Szegedy, C. (2015) Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. International Conference on Machine Learning. Lile, 6-11 July 2015, 448-456. |
[13] | Arima, K., Nagata, F., Shimizu, T., et al. (2023) Improvements of Detection Accuracy and Its Confidence of Defective Areas by YOLOv2 Using a Data Set Augmentation Method. Artificial Life and Robotics, 28, 625-631.
https://doi.org/10.1007/s10015-023-00885-9 |
[14] | 董洪义. 深度学习之PyTorch物体检测实战[M]. 北京: 机械工业出版社, 2020: 171. |
[15] | 闫明艳, 陈根铭, 赖超, 等. 人工智能对肺结核患者病变检出及定性诊断价值研究[J]. 新发传染病电子杂志, 2018, 3(4): 214. |
[16] | 栗俊杰, 毛彭军, 方骞, 等. 基于K210的火点检测和喷头定向控制[J]. 科学技术与工程, 2021, 21(35): 15136-15143. |
[17] | 刘策, 刘小峰. 助老服务机器人多模式友好交互系统设计与实现[J]. 电子测量与仪器学报, 2019(1): 177-182. |