基于VGG-16的子宫颈癌变分级预测  

Prediction of cervical cancer classification based on VGG-16

作　　者：张丽艳 王娟[2] 夏承遗 ZHANG Liyan;WANG Juan;XIA Chengyi(School of Computer Science and Engineering,Tianjin University of Technology,Tianjin 300384,China;School of Electrical Engineering and Automation,Tianjin University of Technology,Tianjin 300384,China)

机构地区：[1]天津理工大学计算机科学与工程学院,天津300384 [2]天津理工大学电气工程与自动化学院,天津300384

出　　处：《天津理工大学学报》2023年第5期21-28,共8页Journal of Tianjin University of Technology

基　　金：国家自然科学基金(61773286,62173247)。

摘　　要：为突破传统人工阅片诊断的局限性,提高对宫颈癌变的筛查效率与准确率,提出一种利用改进后的视觉几何群网络(visual geometry group network,VGG-16)实现女性宫颈病变分级预测的方法,并对原始图像中女性宫颈部位进行感兴趣区域提取及病变位置的定位与分割。在宫颈病变二分类的研究中,通过多次对比试验后,最终测得宫颈病变分级预测的准确率高达92.95%,与未经改进的方法相比,在时间复杂度与空间复杂度上都有明显的降低。试验表明:文中方法不仅能辅助放射科医生进行病变等级诊断,还可提高诊断的效率与准确率,在临床实践中具有重要意义。In order to break through the limitation of traditional manual diagnosis by using film reading and improve the efficiency and accuracy of cervical cancer screening,the improved visual geometry group network is proposed to perform the classification and prediction of female cervical lesions,extract interest regions of female cervix in the original image and locate and segment lesion positions.In the study of the two-classification of cervical lesions,the final accuracy rate of the cervical lesion classification prediction after multiple comparison experiments is raised up to 92.95%.Compared with unimproved methods,the time complexity and the space complexity are significantly reduced.The experimental results show that the proposed method is feasible and this method can not only assist radiologists in the diagnosis of lesion grades,but also improve the efficiency and accuracy of diagnoses.The research is of great significance in clinical practice.

关 键 词：宫颈癌 宫颈病变定位 宫颈病变分割 病变分级预测 感兴趣区域 VGG-16网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]