期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2017
卷号：54
期号：2
页码：91-96
DOI：10.14445/22312803/IJCTT-V54P114
出版社：Seventh Sense Research Group
摘要：A huge type of media Online is creating big challenges for researcher to retrieve a right media for a wright query. This paper describes a suggested formwork to improve retrieving images from aWeb search engine using a dominant meaning method. The dominant meaning is a set of keywords that best fit an intended meaning of a target query. This technique sees the target meaning as a master word and the words that clarify it as slaves. The hypnosis of this research is the more accurate the slaves, the more accurate the retrieving results. The proposed reranking algorithm is built based on the dominant meaning. The contribution of the research is located on the construction of the query and then the filtering of incoming results from theWeb search. The research proposes a calculation method of the rank values forWeb search engine results. The proposed procedures starting with constructing a considerable query from its slaves, sending a query to Our Custom Search Image Engine (as Web Image Search Engine), Retrieve images, retrieving each images document, extracting dominant meaning words from each document included image, calculating rank values of each document based on dominant meaning probability for each document, resorting retrieved images based on the ranking value, and then posting sorted images to a user.
关键词：ensure the proposed system could run successfully; we conducted an empirical analysis on a real example of examination.
