Difference between revisions of "Visual Reranking for Image Retrieval over the Wikipedia Corpus"

From Wikipedia Quality
Jump to: navigation, search
(Visual Reranking for Image Retrieval over the Wikipedia Corpus -- new article)
 
(+ links)
Line 1: Line 1:
'''Visual Reranking for Image Retrieval over the Wikipedia Corpus''' - scientific work related to Wikipedia quality published in 2009, written by Débora Myoupo, Adrian Popescu, Hervé Le Borgne and Pierre-Alain Moëllic.
+
'''Visual Reranking for Image Retrieval over the Wikipedia Corpus''' - scientific work related to [[Wikipedia quality]] published in 2009, written by [[Débora Myoupo]], [[Adrian Popescu]], [[Hervé Le Borgne]] and [[Pierre-Alain Moëllic]].
  
 
== Overview ==
 
== Overview ==
This paper describes the approach authors developed for the WikipediaMM task on 2009 [4], which builds on last year contribution. The main novelties are the refinement of textual query expansion procedure and the introduction of a k-NN based visual reranking procedure. Authors main purpose was to test whether combining textual and content based retrieval improves over purely textual search and the results authors report here confirm that combining modalities results in a significant improvement of results.
+
This paper describes the approach authors developed for the [[Wikipedia]]MM task on 2009 [4], which builds on last year contribution. The main novelties are the refinement of textual query expansion procedure and the introduction of a k-NN based visual reranking procedure. Authors main purpose was to test whether combining textual and content based retrieval improves over purely textual search and the results authors report here confirm that combining modalities results in a significant improvement of results.

Revision as of 14:49, 4 August 2019

Visual Reranking for Image Retrieval over the Wikipedia Corpus - scientific work related to Wikipedia quality published in 2009, written by Débora Myoupo, Adrian Popescu, Hervé Le Borgne and Pierre-Alain Moëllic.

Overview

This paper describes the approach authors developed for the WikipediaMM task on 2009 [4], which builds on last year contribution. The main novelties are the refinement of textual query expansion procedure and the introduction of a k-NN based visual reranking procedure. Authors main purpose was to test whether combining textual and content based retrieval improves over purely textual search and the results authors report here confirm that combining modalities results in a significant improvement of results.