Survey Paper on Content Based Video Retrieval by OCR Technique

Authors: Paridhi Soni, Prof. Abhishek Tiwari

Certificate: View Certificate

Abstract

In the last decade e-lecturing has become more and more popular. The amount of lecture video data on the World Wide Web (WWW) is growing rapidly. Therefore, a more efficient method for video retrieval within large lecture video archives is urgently needed. In this paper the usability and utility study for the video search function in our lecture video portal will be conducted. Automated annotation for OCR results using Linked Open Data resources offers the opportunity to enhance the amount of linked educational resources significantly. Therefore more efficient search and recommendation method could be developed in lecture video archives.

Introduction

Digital video has become a popular and storage medium of exchange because of the rapid development in recording technology, video compression techniques improved and broadband networks in recent years [1]. Therefore, the electuring system is used frequently for audiovisual (audio & video) recordings. An e-lecture consists of slides with relevant points mentioned by the lecturer. A number of colleges and research institutes are taking a chance to record their lectures and publish them online for students to access free of time and location. As a result, there has been a huge increase in the amount of multimedia data on the Web. The user requested for appropriate information which is covered in only few part of the video, and he wants only that information without viewing the complete video. So the problem is how to retrieve the appropriate information in a large lecture video. There are many video search systems like YouTube, Bing etc. based on available textual metadata such as title, person and brief description etc. Text is a high-level semantic feature which has often been used for content-based information retrieval. In lecture videos, texts from lecture slides serve as an outline for the lecture and are very important for understanding. In starting the lecture videos are recorded by a single video camera, which is the cause of lower quality lecture videos. Traditional video retrieval based on feature extraction cannot be efficiently applied to lecture recordings. Lecture recordings are characterized by a homogeneous scene composition. Most of the time, the lecturer is in focus, presenting a topic which is not visible. Thus, image analysis of lecture videos fails even if the producer tries to loosen the scene with creative camera trajectories.

Conclusion

We can conclude as this research work will fulfill some issues by proposed solution. These are the major issues which are not retrieving the video efficiently from the large amount of video database in existing system:- 1. How to retrieve the appropriate information in a large lecture video archive more efficiently. 2. To let machine understand video is important and challenging. 3. How to continuously improve the accuracy ASR lecture video which stills an unsolved problem. 4. The main problem is that the video analysis methods may introduce errors.