LeeDeo: Web-Crawled Academic Video Search Engine

  • Authors:
  • Dongwon Lee;Hung-sik Kim;Eun Kyung Kim;Su Yan;Johnny Chen;Jeongkyu Lee

  • Affiliations:
  • -;-;-;-;-;-

  • Venue:
  • ISM '08 Proceedings of the 2008 Tenth IEEE International Symposium on Multimedia
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present our vision and preliminary design toward web-crawled academic video search engine, named as LeeDeo, that can search, crawl, archive, index, and browse “academic” videos from the Web. Our proposal differs from existing general-purpose search engines such as Google or MSN whose focus is on the search of textual HTML documents or metadata of multimedia objects. Similarly, our proposal also differs from existing academic bibliographic search engines such as CiteSeer, arXiv, or Google Scholar whose focus is on the search and analysis of PDF or PS documents of academic papers. As desiderata of such an academic video search engine, we discuss various issues as follows: (1) Crawling: how to crawl, identify, and download academic videos from the Web? (2) Classification: how to determine the so-called academic videos from the rest? (3) Extraction: how to extract metadata and transcripts from the classified videos? (4) Indexing: how to build indexes for search engines? and (5) Interface: how to provide interface for efficient browse and search of academic videos?