Article Preview
TopIntroduction
Amounts of web data and multimedia content are growing substantially. There are millions of active web pages on current internet according to the different popular search engines. However there are no direct connections between the data on current web pages. Traditional engines for web and manual annotation methods for multimedia are insufficient for accumulated data. The amount of data requires very effective information retrieval systems. Semantic web is new generation of web. Instead of separated and disconnected objects, in semantic web each object is defined with URI and connected with each other by RDF. As a result of RDF structure, each of the objects and their connections can be queried by SPARQL. The semantic search and semantic search engines are new approach to overcome the problem of handling accumulated data.
The semantic concept firstly considered by text retrieval. This was due to its simple form. Then many numbers of papers refer to the demand for new field which is the semantic multimedia retrieval systems (Menemencioğlu & Orak, 2014; Gallego, Corcho, Fernández, Martínez-Prieto, & Suárez-Figueroa, 2013; Esmaili & Abolhassani, 2009). According to the YouTube data over 6 billion hours of video are watched each month and 100 hours of video are uploaded to YouTube in every minute (YouTube). According to the Instagram data 20 billion photos are shared totally and average per day is 60 million photos (Instagram Stats). Starting as a social media for sharing photos, Instagram also enabled users to share videos which are growing every day. This amount of data confirms the need of very effective multimedia retrieval systems.
In this paper we focus on semantic concepts, semantic search, comparison of traditional and semantic search engines, semantic trends, and text and multimedia retrieval. And this paper aims a review of concepts and building a framework for multimedia retrieval system as a result of review.