Content based video retrieval

TL;DR

This paper explores content-based video retrieval by combining multiple visual features like color and texture to improve search accuracy in large video collections, validated through implementation and user testing.

Contribution

It introduces a multi-feature indexing approach for video retrieval, demonstrating enhanced discrimination and search effectiveness over single-feature methods.

Findings

Multi-feature indexing improves retrieval accuracy.

Videos stored in Oracle 9i database facilitate efficient access.

User study confirms effectiveness of combined features.

Abstract

Content based video retrieval is an approach for facilitating the searching and browsing of large image collections over World Wide Web. In this approach, video analysis is conducted on low level visual properties extracted from video frame. We believed that in order to create an effective video retrieval system, visual perception must be taken into account. We conjectured that a technique which employs multiple features for indexing and retrieval would be more effective in the discrimination and search tasks of videos. In order to validate this claim, content based indexing and retrieval systems were implemented using color histogram, various texture features and other approaches. Videos were stored in Oracle 9i Database and a user study measured correctness of response.