基于SVGG16的乳腺肿块图像识别  

Breast mass image recognition based on SVGG16

作　　者：盘安思 徐胜舟[1] 程时宇 佘逸飞 PAN Ansi;XU Shengzhou;CHENG Shiyu;SHE Yifei(College of Computer Science & Hubei Provincial Engineering Research Center for Intelligent Management of Manufacturing Enterprises, South-Central University for Nationalities, Wuhan 430074, China)

机构地区：[1]中南民族大学计算机科学学院&湖北省制造企业智能管理工程技术研究中心,武汉430074

出　　处：《中南民族大学学报（自然科学版）》2021年第4期410-416,共7页Journal of South-Central University for Nationalities：Natural Science Edition

基　　金：湖北省自然科学基金资助项目(2020CFB541);中央高校基本科研业务费专项资金资助项目(CZY19011)。

摘　　要：针对基于深度学习的乳腺X线摄片肿块识别过程中的特征冗余问题,对VGG16进行了改进,减少模型中卷积层和卷积核的个数,提出一种精简的卷积神经网络模型SVGG16,用于感兴趣区域中肿块的识别.同时,为避免网络模型受小样本量限制出现过拟合现象,通过旋转与翻转操作对感兴趣区域进行数据增强.通过实验对网络模型的性能进行评估,结果表明:SVGG16模型的准确率、精确度、敏感度以及F1_score等评价指标分别达到了90.34%、89.87%、88.75%和0.89,明显优于其他已有的卷积神经网络模型,同时其计算效率也明显高于原始VGG16模型.Aiming at the problem of feature redundancy in mammogram mass recognition based on deep learning,a simplified convolutional neural network SVGG16 is proposed by reducing the number of convolutional layers and convolutional kernels of VGG16,and it is used for the recognition of masses in the region of interest.At the same time,to avoid the over-fitting of the network caused by the limitation of small sample size,the data augmentation is carried out by rotation and flipping operations.The performance of the SVGG16 model is evaluated by experiment,the experimental results show that the accuracy,precision,sensitivity and F1_score of the SVGG16 are 90.34%,89.87%,88.75%and 0.89 respectively,which are significantly better than other existing convolutional neural networks,and its computational efficiency is also significantly higher than the original VGG16 model.

关 键 词：乳腺X线摄片 肿块 识别 VGG16模型 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]