摘要
为了实现对处于相近颜色盒中的手术器械准确进行像素级分割,提出基于全卷积神经网络(FCN)的手术器械图像语义分割模型。该模型使用VGG19作为基础模型,添加3层反卷积层,使用Adam优化器优化网络,实现手术器械和手术盒的像素级分割。实验结果表明:该模型能有效实现图像分割,降低特征值数量,减少网络运行时长。
In order to achieve semantic segmentation of surgical instruments accurately in similar color boxes,proposes a semantic segmentation algo.rithm for images based on Fully Convolutional Networks(FCN).This model uses VGG19 as the base model,adds 4 deconvolution layers and uses the Adam optimizer to optimize the network,which achieves pixel-level segmentation of surgical instruments and cassettes.The experimental results show that the model can segment input image effectively and reduce the network running time,while it has less num.ber of feature-values.
作者
郑腾辉
陶青川
ZHENG Teng-hui;TAO Qing-chuan(College of Electronic Information,Sichuan University,Chengdu 610065)
出处
《现代计算机》
2019年第9期80-84,共5页
Modern Computer
