COMPUTER-EXECUTABLE IMAGE SEARCHING METHOD USING NON-PARAMETRIC BAYESIAN SUM-PRODUCT NETWORKS
展开▼
机译:基于非参数贝叶斯和-乘积网络的计算机可执行图像搜索方法
展开▼
页面导航
摘要
著录项
相似文献
摘要
The present invention relates to a computer-executable image searching method through non-parametric Bayesian sum-product networks, comprising the following steps of: receiving a plurality of image data and at least one text data related to each of the image data so as to generate modal information on first image data and at least one first text data as first image modal information and at least one text modal information; performing associated learning through non-parametric Bayesian sum-product network framework with respect to the first image modal information and the at least one first text modal information; searching for text modal information related to a specific text when the specific text is received from a user terminal; and when the search is successful, searching for at least one corresponding image modal information, which has a correlation with corresponding text modal information more than a specific standard so as to provide at least one corresponding image data. Therefore, modal information with respect to data can be learned through the non-parametric Bayesian sum-product networks, thereby providing a quick search service.
展开▼