摘要
介绍了大数据技术中的卷积神经网络和多模态智能技术中的支持向量机、谱聚类等模型。以图像检索和图像分割为例,详述了采用卷积神经网络遴选图像特征表征之后,运用多模态技术进行模型参数自动定参和实际运用的实现过程。该研究项目用于计算机视觉课程的实验教学,有利于学生学习计算机视觉领域的前沿技术,运用不同模型解决实际问题,锻炼学生组织实验、分析实验数据和团队协作等综合能力。
This paper elaborates on convolutional neural network in deeply learning as well as support vector machine and spectral clustering in intelligent multi-modality techniques.Two types of experiments including image retrieval and image segmentation are described.The strategy of adopting deep learning for feature extraction as well as intelligent multi-modality techniques for parameters tuning is introduced in detail.This design of novel experiments is specialized in the Computer Vision course,which is beneficial for the students to get familiar with up-to-date knowledge in computer vision,know how to solve practical problems via diverse models,and learn the capability to organize experiments,analyze data as well as team collaboration in experiments.
出处
《实验技术与管理》
CAS
北大核心
2016年第9期122-125,共4页
Experimental Technology and Management
基金
国家自然科学基金项目(61403182
61363046)
江西省教育厅高校教改项目(JXJG-15-1-26)
江西省青年科学家培养对象项目(20153BCB23029)
教育部留学回国人员科研启动基金项目([2014]1685)
关键词
计算机视觉
大数据
多模态
实验教学
computer vision
big data
multi-modality
experimental teaching
