Content-Based Video Retrieval Based on Integration of Wavelet Transform, Color and Texture Features

Abstract

High-resolution, large-sized videos can now be transferred due to the fast developmentof information and communication technology, and video applicationshave developed in line with data quality levels. The applications of contentbasedvideo retrieval (CBVR) in various fields, like surveillance, education, sports,medicine etc., make it a crucial video application. The efficient Content-BasedVideo Retrieval (CBVR) algorithm described in this thesis is based on the MPEG-7 features of the Discrete Wavelet Transform (DWT), Dual-Tree Complex WaveletTransform (DTCWT), Edge Histogram Descriptor (EHD), Linear Binary Pattern(LBP) and colour Layout Descriptor (CLD).This thesis proposes a content-based video retrieval system that integrates wavelettransform, colour, and texture features for efficient and accurate video retrieval.The proposed system aims to address the limitations of traditional video retrievalsystems that rely solely on low-level features, such as colour and texture, and donot consider the video�s structural information. The system utilizes the wavelettransform to extract the structural information of the video. It combines it withcolour and texture features to create a robust and accurate feature set for retrieval.