Mixed ranking scheme for video retrieval

Y. Feng, Jinchang Ren, J. Jiang

Research output: Contribution to journalArticlepeer-review


A unified ranking scheme for effective video retrieval is proposed, in which low-level visual feature terms and high-level image category features are combined organically to inspire effective retrieval in the manner of semantics. By taking these features as a joint fact of document relevance, the BM25 model, popular in text retrieval, is employed to determine a mixed similarity rank of video documents. Experiments using the well-known TRECVID retrieval dataset have validated the superiority of the methodology.
Original languageEnglish
Pages (from-to)1600-1601
Number of pages2
JournalElectronics Letters
Issue number24
Publication statusPublished - 25 Nov 2010


  • video retrieval
  • feature extraction
  • content-based retrieval
  • mixed ranking scheme
  • video document ranking

Fingerprint Dive into the research topics of 'Mixed ranking scheme for video retrieval'. Together they form a unique fingerprint.

Cite this