Adaptive Multimedia Retrieval: User, Context, and Feedback: - download pdf or read online

By Myunggwon Hwang, Hyunjang Kong, Sunkyoung Baek, Pankoo Kim (auth.), Stéphane Marchand-Maillet, Eric Bruno, Andreas Nürnberger, Marcin Detyniecki (eds.)

ISBN-10: 3540715444

ISBN-13: 9783540715443

ISBN-10: 3540715452

ISBN-13: 9783540715450

This ebook constitutes the completely refereed post-proceedings of the 4th foreign Workshop on Adaptive Multimedia Retrieval, AMR 2006, held in Geneva, Switzerland in July 2006. The 18 revised complete papers provided including 2 invited papers have been conscientiously chosen in the course of rounds of reviewing and development. additionally integrated are invited contributions which were meant to open on less-addressed themes in the neighborhood, because it is the case for song details retrieval and disbursed info retrieval. The papers are prepared in topical sections on ontology-based retrieval and annotation, score and similarity measurements, track info retrieval, visible modelling, adaptive retrieval, structuring multimedia, in addition to person integration and profiling.

Show description

Read or Download Adaptive Multimedia Retrieval: User, Context, and Feedback: 4th International Workshop, AMR 2006, Geneva, Switzerland, July 27-28, 2006, Revised Selected Papers PDF

Similar computers books

Sphinx Search: Beginner's Guide - download pdf or read online

This ebook is a step by step advisor for an absolute newbie. It covers every thing, from fitting to configuring, to get you all started quick. It has a number of code examples that the reader can attempt on their lonesome and research whereas doing so. It has full-fledged purposes as examples that readers can stick to.

Download PDF by Ricardo Baeza-Yates (auth.), José R. Correa, Alejandro: LATIN 2006: Theoretical Informatics: 7th Latin American

This booklet constitutes the refereed court cases of the seventh foreign Symposium, Latin American Theoretical Informatics, LATIN 2006, held in Valdivia, Chile in March 2006. The sixty six revised complete papers offered including seven invited papers have been conscientiously reviewed and chosen from 224 submissions.

Computer Music Modeling and Retrieval: International by Rubén Hinojosa Chapel (auth.), Uffe Kock Wiil (eds.) PDF

This ebook constitutes the completely refereed post-proceedings of the foreign laptop track Modeling and Retrieval Symposium, CMMR 2003, held in Montpellier, France, in may possibly 2003. The 20 revised complete papers have been conscientiously chosen in the course of rounds of reviewing and development. a result of interdisciplinary nature of the world, the papers handle a extensive number of subject matters together with info retrieval, programming, human-computer interplay, electronic libraries, hypermedia, man made intelligence, acoustics, sign processing, and so forth.

Kenneth H. Buetow (auth.), Sarah Cohen-Boulakia, Val Tannen's Data Integration in the Life Sciences: 4th International PDF

Figuring out the mechanisms curious about existence (e. g. , studying the organic functionofasetofproteins,inferringtheevolutionofasetofspecies)isbecoming increasinglydependent onprogressmade inmathematics,computer science,and molecular engineering. For the earlier 30 years, new high-throughput applied sciences were built producing quite a lot of info, dispensed throughout many info resources on the internet, with a excessive measure of semantic heterogeneity and di?

Additional resources for Adaptive Multimedia Retrieval: User, Context, and Feedback: 4th International Workshop, AMR 2006, Geneva, Switzerland, July 27-28, 2006, Revised Selected Papers

Example text

We use a semi-automatic object identification algorithm to identify objects. Either by experts or by using a training set, the properties of objects and relations between objects are determined. Object classification is done by: – Low-level feature values like color distribution, shape, texture. – Spatial knowledge. – Spatio-temporal change. (A set of consecutive key frames must be searched to gain knowledge for this property) Extraction phase starts when we have enough knowledge to separate one object from others.

Given a query q, a retrieval system then simply computes the scores {F (q, p), ∀p ∈ P } and ranks the pictures of P by decreasing scores. The effectiveness of such a system is hence mainly determined by the choice of an appropriate function F . In fact, optimal retrieval performance would be achieved if F satisfies / R(q), F (q, p+ ) > F (q, p− ), ∀q, ∀p+ ∈ R(q), ∀p− ∈ (1) where R(q) refers to the pictures of P which are relevant to q. In other words, if F satisfies (1), the retrieval system will always rank the relevant pictures above the non-relevant ones.

Such models include, for instance, CrossMedia Relevance Models (CMRM) [3], Latent Dirichlet Allocation (LDA) [5] or Probabilistic Latent Semantic Analysis (PLSA) [6]. In this paper, we introduce an alternative to these approaches. The proposed model, Passive-Aggressive Model for Image Retrieval (PAMIR), relies on discriminative learning. This means that the model parameters are not selected to maximize the likelihood of some annotated training data; they are instead selected to maximize the retrieval performance of the model over a set of training queries.

Download PDF sample

Adaptive Multimedia Retrieval: User, Context, and Feedback: 4th International Workshop, AMR 2006, Geneva, Switzerland, July 27-28, 2006, Revised Selected Papers by Myunggwon Hwang, Hyunjang Kong, Sunkyoung Baek, Pankoo Kim (auth.), Stéphane Marchand-Maillet, Eric Bruno, Andreas Nürnberger, Marcin Detyniecki (eds.)


by Kevin
4.4

Rated 4.85 of 5 – based on 34 votes