Graduate School and Research Center in Digital Sciences

Linking text and visual concepts semantically for cross modal multimedia search

Safadi, Bahjat; Sahuguet, Mathilde; Huet, Benoit

ICIP 2014, 21st IEEE International Conference on Image Processing, October 27-30, 2014, Paris, France

Currently, popular search engines retrieve documents on the basis of text information. However, integrating the visual information with the text-based search for video and image retrieval is still a hot research topic. In this paper, we propose and evaluate a video search framework based on using visual information to enrich the classic text-based search for video retrieval. With the proposed framework, we endeavor to show experimentally, on a set of real world scenarios, that visual cues can effectively contribute to significant quality improvement of video retrieval. Experimental results show that mapping text-based queries to visual concepts improves the performance of the search system. Moreover, when appropriately selecting the relevant visual concepts for a query, a very substantial improvement of the system's performance is achieved.

Document Doi Bibtex

Title:Linking text and visual concepts semantically for cross modal multimedia search
Keywords:Multimedia retrieval, video search and visual cues
Department:Data Science
Eurecom ref:4316
Copyright: © 2014 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
Bibtex: @inproceedings{EURECOM+4316, doi = {}, year = {2014}, title = {{L}inking text and visual concepts semantically for cross modal multimedia search}, author = {{S}afadi, {B}ahjat and {S}ahuguet, {M}athilde and {H}uet, {B}enoit}, booktitle = {{ICIP} 2014, 21st {IEEE} {I}nternational {C}onference on {I}mage {P}rocessing, {O}ctober 27-30, 2014, {P}aris, {F}rance}, address = {{P}aris, {FRANCE}}, month = {10}, url = {} }
See also: