Learning for Meta-Recognition

dc.creatorScheirer, WJ
dc.creatorRocha, AD
dc.creatorParris, J
dc.creatorBoult, TE
dc.date2012
dc.dateAUG
dc.date2014-08-01T18:33:55Z
dc.date2015-11-26T17:05:59Z
dc.date2014-08-01T18:33:55Z
dc.date2015-11-26T17:05:59Z
dc.date.accessioned2018-03-28T23:54:23Z
dc.date.available2018-03-28T23:54:23Z
dc.identifierIeee Transactions On Information Forensics And Security. Ieee-inst Electrical Electronics Engineers Inc, v. 7, n. 4, n. 1214, n. 1224, 2012.
dc.identifier1556-6013
dc.identifierWOS:000306520900010
dc.identifier10.1109/TIFS.2012.2192430
dc.identifierhttp://www.repositorio.unicamp.br/jspui/handle/REPOSIP/80821
dc.identifierhttp://repositorio.unicamp.br/jspui/handle/REPOSIP/80821
dc.identifier.urihttp://repositorioslatinoamericanos.uchile.cl/handle/2250/1279734
dc.descriptionFundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)
dc.descriptionIn this paper, we consider meta-recognition, an approach for postrecognition score analysis, whereby a prediction of matching accuracy is made from an examination of the tail of the scores produced by a recognition algorithm. This is a general approach that can be applied to any recognition algorithm producing distance or similarity scores. In practice, meta-recognition can be implemented in two different ways: a statistical fitting algorithm based on the extreme value theory, and a machine learning algorithm utilizing features computed from the raw scores. While the statistical algorithm establishes a strong theoretical basis for meta-recognition, the machine learning algorithm is more accurate in its predictions in all of our assessments. In this paper, we present a study of the machine learning algorithm and its associated features for the purpose of building a highly accurate meta-recognition system for security and surveillance applications. Through the use of feature-and decision-level fusion, we achieve levels of accuracy well beyond those of the statistical algorithm, as well as the popular 'cohort' model for postrecognition score analysis. In addition, we also explore the theoretical question of why machine learning-based algorithms tend to outperform statistical meta-recognition and provide a partial explanation. We show that our proposed methods are effective for a variety of different recognition applications across security and forensics-oriented computer vision, including biometrics, object recognition, and content-based image retrieval.
dc.description7
dc.description4
dc.description1214
dc.description1224
dc.descriptionONR STTR [N00014-07-M-0421]
dc.descriptionONR SBIR [N00014-09-M-0448]
dc.descriptionONR MURI [N00014-08-1-0638]
dc.descriptionDHS SBIR [NBCHC080054]
dc.descriptionFundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)
dc.descriptionMicrosoft
dc.descriptionFundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)
dc.descriptionONR STTR [N00014-07-M-0421]
dc.descriptionONR SBIR [N00014-09-M-0448]
dc.descriptionONR MURI [N00014-08-1-0638]
dc.descriptionDHS SBIR [NBCHC080054]
dc.descriptionFAPESP [2010/05647-4]
dc.languageen
dc.publisherIeee-inst Electrical Electronics Engineers Inc
dc.publisherPiscataway
dc.publisherEUA
dc.relationIeee Transactions On Information Forensics And Security
dc.relationIEEE Trans. Inf. Forensic Secur.
dc.rightsfechado
dc.rightshttp://www.ieee.org/publications_standards/publications/rights/rights_policies.html
dc.sourceWeb of Science
dc.subjectContent-based image retrieval
dc.subjectface recognition
dc.subjectfingerprint recognition
dc.subjectmachine learning
dc.subjectmeta-recognition
dc.subjectmultibiometric fusion
dc.subjectobject recognition
dc.subjectperformance modeling
dc.subjectsimilarity scores
dc.subjectFace Recognition
dc.subjectPerformance
dc.titleLearning for Meta-Recognition
dc.typeArtículos de revistas


Este ítem pertenece a la siguiente institución