摘要
图像分类是激光技术在医学领域中的一个重要应用,针对大规模激光医学图像分类效率、正确率低的缺陷,提出一种基于云计算的激光医学图像分类方法。首先对激光医学图像进行预处理,增强图像质量,然后提取激光医学图像的特征,并采用极限学习机对特征向量进行训练,提高激光医学图像分类正确率,最后采用云计算的Map/Reduce对激光医学图像分类过程进行并行处理,提高激光医学图像分类的效率。采用激光医学图像数据实现仿真实验,结果表明,本文方法的激光医学图像分类正确率要优于传统激光医学图像分类方法,而且具有优异并行运行性能,激光医学图像分类速度明显优于传统方法,具有一定的实际应用价值。
Image classification is an important application of laser technology in medical field. A laser medical image classification method based on cloud computing is proposed to solve the defects of poor classification efficiency and low accuracy. Firstly,the laser medical image is pre-processed to enhance the image quality,and then the characteristics of the laser medical image are extracted which is trained by the extreme learning machine to improve the accuracy of laser medical image classification,finally,the Map / Reduce of cloud computing is used to deal with the process of laser medical image classification to improve the efficiency of laser medical image classification. The simulation experiment was realized by using laser medical image data,the results show that the classification accuracy of laser medical image is better than that of traditional method,and has excellent parallel performance,laser medical image classification rate is significantly better than the traditional method,it has certain practical application value.
出处
《激光杂志》
北大核心
2016年第4期54-57,共4页
Laser Journal
基金
国家自然科学基金面上项目(NSF61472051)
重庆市科委前沿与应用基础研究项目(cstc2014jcyj A40010)
重庆市教委科学技术研究项目(KJ111604)
关键词
激光成像技术
图像分类
云计算
分类方法
特征提取
laser imaging technology
image classification
cloud computing
classification method
feature extraction
