Skip navigation
Skip navigation

Incremental Training of a Detector Using Online Sparse Eigendecomposition

Paisitkriangkrai, Sakrapee; Shen, Chunhua; Zhang, Jian

Description

The ability to efficiently and accurately detect objects plays a very crucial role for many computer vision tasks. Recently, offline object detectors have shown a tremendous success. However, one major drawback of offline techniques is that a complete set of training data has to be collected beforehand. In addition, once learned, an offline detector cannot make use of newly arriving data. To alleviate these drawbacks, online learning has been adopted with the following objectives: 1) the...[Show more]

dc.contributor.authorPaisitkriangkrai, Sakrapee
dc.contributor.authorShen, Chunhua
dc.contributor.authorZhang, Jian
dc.date.accessioned2015-12-10T23:22:22Z
dc.identifier.issn1057-7149
dc.identifier.urihttp://hdl.handle.net/1885/66490
dc.description.abstractThe ability to efficiently and accurately detect objects plays a very crucial role for many computer vision tasks. Recently, offline object detectors have shown a tremendous success. However, one major drawback of offline techniques is that a complete set of training data has to be collected beforehand. In addition, once learned, an offline detector cannot make use of newly arriving data. To alleviate these drawbacks, online learning has been adopted with the following objectives: 1) the technique should be computationally and storage efficient; 2) the updated classifier must maintain its high classification accuracy. In this paper, we propose an effective and efficient framework for learning an adaptive online greedy sparse linear discriminant analysis model. Unlike many existing online boosting detectors, which usually apply exponential or logistic loss, our online algorithm makes use of linear discriminant analysis' learning criterion that not only aims to maximize the class-separation criterion but also incorporates the asymmetrical property of training data distributions. We provide a better alternative for online boosting algorithms in the context of training a visual object detector. We demonstrate the robustness and efficiency of our methods on handwritten digit and face data sets. Our results confirm that object detection tasks benefit significantly when trained in an online manner.
dc.publisherInstitute of Electrical and Electronics Engineers (IEEE Inc)
dc.sourceIEEE Transactions on Image Processing
dc.subjectKeywords: Asymmetry; Classification accuracy; Eigen decomposition; Face data; Feature selection; Handwritten digit; Incremental training; Learning criterion; Linear discriminant analysis; object detection; Object detectors; Offline; On-line algorithms; On-line boos Asymmetry; feature selection; greedy sparse linear discriminant analysis (GSLDA); object detection; online linear discriminant analysis
dc.titleIncremental Training of a Detector Using Online Sparse Eigendecomposition
dc.typeJournal article
local.description.notesImported from ARIES
local.identifier.citationvolume20
dc.date.issued2011
local.identifier.absfor080106 - Image Processing
local.identifier.ariespublicationf2965xPUB1293
local.type.statusPublished Version
local.contributor.affiliationPaisitkriangkrai, Sakrapee, University of Adelaide
local.contributor.affiliationShen, Chunhua, College of Engineering and Computer Science, ANU
local.contributor.affiliationZhang, Jian, NICTA
local.description.embargo2037-12-31
local.bibliographicCitation.issue1
local.bibliographicCitation.startpage213
local.bibliographicCitation.lastpage226
local.identifier.doi10.1109/TIP.2010.2053548
local.identifier.absseo970108 - Expanding Knowledge in the Information and Computing Sciences
dc.date.updated2016-02-24T08:12:07Z
local.identifier.scopusID2-s2.0-79551532470
CollectionsANU Research Publications

Download

File Description SizeFormat Image
01_Paisitkriangkrai_Incremental_Training_of_a_2011.pdf3.85 MBAdobe PDFThumbnail


Items in Open Research are protected by copyright, with all rights reserved, unless otherwise indicated.

Updated:  19 May 2020/ Responsible Officer:  University Librarian/ Page Contact:  Library Systems & Web Coordinator