王兆平,陈炳坤.基于社交网络元数据的图像分类算法[J].深圳大学学报理工版,2019,36(No.4(347-472)):453-459.[doi:10.3724/SP.J.1249.2019.04453]
WANG Zhaoping and CHEN Bingkun.Image classification algorithm based on social network metadata[J].Journal of Shenzhen University Science and Engineering,2019,36(No.4(347-472)):453-459.[doi:10.3724/SP.J.1249.2019.04453]
点击复制
基于社交网络元数据的图像分类算法
王兆平1,陈炳坤2

1)长沙民政职业技术学院,湖南长沙 410004; 2)深圳大学计算机与软件学院,广东深圳 518060

人工智能; 社交网络; 网络表征学习; 图像分类; 神经网络; 社交多媒体

Image classification algorithm based on social network metadata
WANG Zhaoping1 and CHEN Bingkun2

1)Changsha Social Work College, Changsha 410004, Hunan Province, P.R.China2)College of Computer Science and Software Engineering Shenzhen University, Shenzhen 518060, Guangdong Province, P.R.China

artificial intelligence; social network; network representation learning; image classification; neural network; social multimedia

DOI: 10.3724/SP.J.1249.2019.04453

备注

在线社交网络图像通常携带大量的社交网络元数据,包含了丰富的图像语义信息,可以帮助用户区分图片中的内容.提出一种基于社交网络元数据的图像分类(multiple social metadata image classification networks,简称MSNet)算法,首先采集得到图像的多种社交网络元数据,根据图像社交网络信息构造出图像的多种关系网络,然后使用网络表征学习算法学习出图像在各个关系网络中的表征向量,最后使用图像的视觉特征和网络表征训练一个神经网络分类器对图像进行分类.通过在PASCAL、MIR、CLEF和NUS数据集上对比MSNet与CNN-neighbor、核典型相关分析(kernel canonical correlation analysis, KCCA)算法的性能,证明了MSNet算法能提升图像分类的性能.

The images in online social networks usually carry a large amount of social network metadata, such as labels, users, picture groups, locations and comments. These social network metadata, which include rich image semantic information, can help users distinguish the content of imagesand communicate with each other. An image classification algorithm based on multiple social networks, MSNet, is proposed in this paper. Firstly, MSNet collects the social network metadata of images and constructs relationship networks of images based on the metadata. Then,a network embedding algorithm is used to learn the representation vectors of images in each network. Finally,a neural network classifier is trained to classify the images by using the visual features and network representation of images. The experimental results on PASCAL、MIR、CLEF and NUS image data sets show the superiority performance of MSNet in comparison with CNN-neighbor and kernel canonical correlation analysis(KCCA).

·