| Title | Optimizing Visual Vocabularies Using Soft Assignment Entropies |
| Authors | Yubin Kuang, Karl Åström, Lars Kopp, Magnus Oskarsson, Martin Byröd |
| Full-text | Available as PDF |
| Alternative Location | http://dx.doi.org/10.1007/9..., Restricted Access |
| Publication | Lecture Notes in Computer Science |
| Year | 2011 |
| Volume | 6495 |
| Pages | 255 - 268 |
| Document type | Conference paper |
| Conference name | 10th Asian Conference on Computer Vision (ACCV 2010) |
| Conference Date | 2010-11-08/2010-11-12 |
| Conference Location | Queenstown, New Zealand |
| Status | Published |
| Quality controlled | Yes |
| Language | eng |
| Publisher | Springer |
| Abstract English | The state of the art for large database object retrieval in images is based on quantizing descriptors of interest points into visual words. High similarity between matching image representations (as bags of words) is based upon the assumption that matched points in the two images end up in similar words in hard assignment or in similar representations in soft assignment techniques. In this paper we study how ground truth correspondences can be used to generate better visual vocabularies. Matching of image patches can be done e.g. using deformable models or from estimating 3D geometry. For optimization of the vocabulary, we propose minimizing the entropies of soft assignment of points. We base our clustering on hierarchical k-splits. The results from our entropy based clustering are compared with hierarchical k-means. The vocabularies have been tested on real data with decreased entropy and increased true positive rate, as well as better retrieval performance. |
| ISBN/ISSN/Other | ISSN: 1611-3349 (online) ISBN: 978-3-642-19282-1(online) |
Questions: webmaster
Last update: 2013-04-11
Centre for Mathematical Sciences, Box 118, SE-22100, Lund. Telefon: +46 46-222 00 00 (vx)