Graduate School and Research Center In communication systems

Multimedia indexing and retrieval

[MMIR]
T Technical Teaching


Abstract

  • While search engines such as Google are extremely efficient for huge databases of text documents, searching for image and video information is still a research challenge.
  • The objective of this course is to study some of the problems and techniques that are involved in the construction of multimedia search engines.

Description

  • This course will start with an introduction to the basics of information retrieval, the construction and structure of index files, the major similarity models, and the evaluation of search performance. It will describe the specificities of web searching, such as Google's PageRank.
  • Then, it will explore image processing techniques, which are able to process automatically image and video to extract content descriptors, in particular, video shot segmentation, feature extraction (color, texture, and movement), text recognition, video classification, etc.
  • Finally, the course will also cover the MPEG-7 standard for encoding multimedia content description and present some applications of multimedia search such as video browsers and personalized filtering engines.
Nb hours: 21.00
Control form: examen écrit