Issues and challenges for content-based video search engines a survey

Abstract

With the vast growth and progress of multimedia on the Internet, particularly videos, which caused a growing demand for video retrieval, organizing, and automate systems, as many users mandate content-based retrieval systems over text-based retrieval systems. The process of extracting and selecting features, plays a very important role in content-based video retrieval regardless the significance of video characteristics. The reduction of time and space costs of the retrieval process is dependent on effective features selection. In this paper, we will discuss content-based video indexing and retrieval, in the application of video search engines CBVSE, which is the problem of searching for digital videos over the Internet. Addressing diverse approaches regarding video indexing and retrieval with brief outline and classification. Furthermore, we will propose a framework for a CBVSE that will offer a less complex solution for content-based video retrieval for video search engines over the web using reduced features vector with better accuracy to decrease the complexity trade-off. These features include temporal boundaries patterns combination and key-object concept features, as well as classifying the index list with a multi key-object concepts arrangement built on the statistical redundancy of the detected key-object concepts for each video shot in the sequence.