Hunting moving targets: an extension to {B}ayesian methods in multimediadatabases

Bibtex entry :

@inproceedings { VG:MSM1999b,
    author = { Wolfgang M{\"u}ller and David McG. Squire and Henning M{\"u}ller and Thierry Pun },
    title = { Hunting moving targets: an extension to {B}ayesian methods in multimediadatabases },
    booktitle = { Multimedia Storage and Archiving Systems IV (VV02) },
    year = { 1999 },
    volume = { 3846 },
    series = { SPIE Proceedings },
    address = { Boston, Massachusetts, USA },
    month = { 20--22~September },
    keywords = { relevance feedback, query drift, target testing, Bayesian methods,user modelling },
    note = { (SPIE Symposium on Voice, Video and Data Communications) },
    url = { },
    abstract = { It has been widely recognised that the difference between the levelof abstraction of the formulation of a query (by example) and thatof the desired result (usually an image with certain semantics) callsfor the use of learning methods that try to bridge this gap. Cox\emph{et al.}~have proposed a Bayesian method to learn the user'spreferences during each query. Cox \emph{et al.}\'s system, \texttt{PicHunter},is designed for optimal performance when the user is searching fora fixed target image. The performance of the system was evaluatedusing target testing, which ranks systems according to the numberof interaction steps required to find the target, leading to simple,easily reproducible experiments. There are some aspects of imageretrieval, however, which are not captured by this measure. In particular,the possibility of query drift (i.e.~a moving target) is completelyignored. The algorithm proposed by Cox \emph{et al.}~does not copewell with a change of target at a late query stage, because it isassumed that user feedback is noisy, but consistent. In the caseof a moving target, however, the feedback is noisy \emph{and} inconsistentwith earlier feedback. In this paper we propose an enhanced Bayesianscheme which selectively forgets inconsistent user feedback, thusenabling both the program and the user to ``change their minds''.The effectiveness of this scheme is demonstrated in moving targettests on a database of heterogeneous real-world images. },
    owner = { steph },
    timestamp = { 2008.05.04 },
    url1 = { },
    vgclass = { refpap },
    vgproject = { viper },

Keywords: machine learning, information geometry, data mining, Big Data, affective information retrieval (recherche d'information), information visualisation, content-based image and video retrieval (CBIR, CBR, CBVR, CBMR, CBMIR), information mining, classification, multimedia and multimodal information management, semantic web, knowledge base (RDF, OWL, XML, metadata, auto-annotation, description), multimodal information fusion