Content Based Video Retrieval System

Content Based Video Retrieval System is a Desktop Application that allows to create multimodal index of video dataset(i.e. Textual index, Visual Index, Acoustic Index) and allow the user to search through that index using multimodality, i.e. user can search using textual, visual or acoustic query. S

Project Title

Content Based Video Retrieval System

Project Area of Specialization

Artificial Intelligence

Project Summary

Content Based Video Retrieval System is a Desktop Application that allows to create multimodal index of video dataset(i.e. Textual index, Visual Index, Acoustic Index) and allow the user to search through that index using multimodality, i.e. user can search using textual, visual or acoustic query. System will retrieve the results for entered query from index (i.e. for textual query, results retrieved form textual index) and rank them. In case of Multimodal query( i.e. more than one query option selected), System will retrieve results from multiple indexes, Merge them and then give Ranked Results.

Project Objectives

The main objectives of this project are:

• Development of an interactive system which provides users with a simple and convenient interface for searching in video dataset using various methods.
• It will provide multimodal query search (textual, Acoustic, Visual) for more effective video retrieval.
• It will help users to retrieve desired video from a video index efficiently based on the video contents
• It will merge the results in case of multi-query search i.e. user searches using different search methods
• It provides ranked results.

Project Implementation Method

Tools and Techniques:

1. Eclipse

2. Matlab R2018a

3. Lire, Lucene and jMir libraries for indexing

Method:

In Content Based Video Retrieval System. Lucene library is used for textual indexing and retrieval, Lire for Visual Indexing and retrieval and JMir for audio.

In part 1,  Text, Keyframes and audio will be seperated from videos and seperate indexes are created for each type. 

In part 2, Front end i.e. interface for application is created for user interaction.

In part3,  Merging and Ranking will be done.

Benefits of the Project

• This project provides efficient Retrieval over large video dataset.
• Facilitates the user by providing functionality of multimodal search as mostly existing system rely on textual search only.

Technical Details of Final Deliverable

Desktop Application that loads a Video Dataset to index,  Retrieves from the indexed dataset based on multimodal query, Merge and Ranks Results.

Final Deliverable of the Project

Software System

Type of Industry

IT

Technologies

Artificial Intelligence(AI), Big Data

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com