Image Retrieval using Collaborative Filtering and Visual Navigation

ACM Special Interest Group on Computer Graphics and Interactive Techniques Conference (SIGGRAPH '10)

Authors: Kai Uwe Barthel, Sebastian Müller, David Backstein, Dirk Neumann, Klaus Jung

TL;DR

We propose a new image search system that combines keywords, low-level visual features, and collaborative filtering to create a network of semantic relationships between images. Unlike other methods, our system models image similarity through a network of linked images, with link weights learned solely from user interactions.

Abstract

Internet image search systems mostly use words from the context of the web page containing the image as keywords. The performance of these search systems is rather poor, as the search systems neither know the intention of the searching user nor the semantic relationships of these images. Content-based image retrieval (CBIR) systems rely on the assumption that similar images share similar visual features. Despite intense research efforts, the results of CBIR systems have not reached the performance of text based search engines. The main problem of CBIR systems is the semantic gap between the content that can be described with low-level visual features and the description of image content that humans use with high-level semantic concepts. Some image retrieval systems have combined the keyword and the content-based visual search approach. However with this approach many images may be found that semantically do not match. In addition semantically similar images that visually look different cannot be found at all.

BibTeX

If you use our work in your research, please cite our publication:

@inproceedings{10.1145/1836845.1836939,
author = {Barthel, Kai Uwe and M\"{u}ller, Sebastian and Backstein, David and Neumann, Dirk and Jung, Klaus},
title = {Image retrieval using collaborative filtering and visual navigation},
year = {2010},
isbn = {9781450303934},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/1836845.1836939},
doi = {10.1145/1836845.1836939},
booktitle = {ACM SIGGRAPH 2010 Posters},
articleno = {87},
numpages = {1},
keywords = {CBIR, collaborative techniques, semantic modeling},
location = {Los Angeles, California},
series = {SIGGRAPH '10}
}