תקציר
We propose a new learning approach for image retrieval, which we call adjustment learning, and demonstrate its use for face recognition and color matching. Our approach is motivated by a frequently encountered problem, namely, that variability in the original data representation which is not relevant to the task may interfere with retrieval and make it very difficult. Our key observation is that in real applications of image retrieval, data sometimes comes in small chunks - small subsets of images that come from the same (but unknown) class. This is the case, for example, when a query is presented via a short video clip. We call these groups chunklets, and we call the paradigm which uses chunklets for unsupervised learning adjustment learning. Within this paradigm we propose a linear scheme, which we call Relevant Component Analysis; this scheme uses the information in such chunklets to reduce irrelevant variability in the data while amplifying relevant variability. We provide results using our method on two problems: face recognition (using a database publicly available on the web), and visual surveillance (using our own data). In the latter application chunklets are obtained automatically from the data without the need of supervision.
שפה מקורית | אנגלית |
---|---|
כותר פרסום המארח | Computer Vision - ECCV 2002 - 7th European Conference on Computer Vision, Proceedings |
עורכים | Anders Heyden, Gunnar Sparr, Mads Nielsen, Peter Johansen |
מוציא לאור | Springer Verlag |
עמודים | 776-790 |
מספר עמודים | 15 |
כרך | 2353 |
מסת"ב (אלקטרוני) | 9783540437482 |
מזהי עצם דיגיטלי (DOIs) | |
סטטוס פרסום | פורסם - 2002 |
פורסם באופן חיצוני | כן |
אירוע | 7th European Conference on Computer Vision, ECCV 2002 - Copenhagen, דנמרק משך הזמן: 28 מאי 2002 → 31 מאי 2002 |
סדרות פרסומים
שם | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|
כרך | 2353 |
ISSN (מודפס) | 0302-9743 |
ISSN (אלקטרוני) | 1611-3349 |
כנס
כנס | 7th European Conference on Computer Vision, ECCV 2002 |
---|---|
מדינה/אזור | דנמרק |
עיר | Copenhagen |
תקופה | 28/05/02 → 31/05/02 |
הערה ביבליוגרפית
Publisher Copyright:© Springer-Verlag Berlin Heidelberg 2002.