Multimedia database search techniques

Abstract

From the past decade, the database area has been quite energetic. Several efficient methods are being discovered for managing alphanumeric data. However, when new types of data in the form of image, video and audio are first brought into a database environment, these methods failed to efficiently manage the new kinds of data. Therefore, new multimedia indexing and retrieval methods are required. This project seeks to investigate possible approaches for deriving multimedia search techniques. The project scope can be categorized into four areas, understanding of different media types, extensive investigation of multimedia database management system, broad research into multimedia document retrieval and implementing some available methods for audio and image retrieval using MATLAB. In this project, LPC analysis and Filter-Bank analysis are implemented for speaker recognition. In addition, Histogram based image retrieval technique is implemented for color image retrieval. Results are presented for these methods and their performance is discussed.