Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

A novel incremental principal component analysis and its application for face recognition.

Haitao Zhao1, Pong Chi Yuen, James T Kwok

  • 1Institute of Aerospace Science and Technology, Shanghai Jiaotong University, China. zhaoht@sjtu.edu.cn

IEEE Transactions on Systems, Man, and Cybernetics. Part B, Cybernetics : a Publication of the IEEE Systems, Man, and Cybernetics Society
|August 15, 2006
PubMed
Summary

A new incremental Principal Component Analysis (IPCA) method, SVDU-IPCA, addresses scaling issues in face recognition. It bounds approximation errors, offering a scalable alternative to traditional PCA methods.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Development and evaluation of <sup>18</sup>F-labeled novel radiopharmaceuticals for PET imaging of fibroblast activation protein expressing tumors.

Bioorganic chemistry·2025
Same author

Preliminary Observations of a Substrate-Based Radiotracer Biosensor for <i>In Vivo</i> Positron Emission Tomography Imaging of Tumor Transmembrane Protease ST14.

ACS sensors·2025
Same author

Mechano-Chemo-Electrochemically Booming Nickel-Rich Layered Cathode Electrochemical Performance.

ACS applied materials & interfaces·2025
Same author

Combined TACE with Targeted and Immunotherapy versus TACE Alone Improves DFS in HCC with MVI: A Multicenter Propensity Score Matching Study.

Journal of hepatocellular carcinoma·2025
Same author

Highly Accurate Adaptive Federated Forests Based on Resistance to Adversarial Attacks in Wireless Traffic Prediction.

Sensors (Basel, Switzerland)·2025
Same author

A Knowledge-Guided Graph Learning Approach Bridging Phenotype- and Target-Based Drug Discovery.

Advanced science (Weinheim, Baden-Wurttemberg, Germany)·2025

Area of Science:

  • Computer Science
  • Machine Learning
  • Pattern Recognition

Background:

  • Principal Component Analysis (PCA) is effective for pattern recognition and image analysis, notably in face recognition algorithms like eigenface and fisherface.
  • Existing PCA-based face recognition systems face scalability challenges due to high computational and memory demands.
  • Incremental PCA (IPCA) methods offer a solution but often lack guaranteed approximation error bounds.

Purpose of the Study:

  • To propose a novel Incremental PCA (IPCA) method, Singular Value Decomposition updating-based IPCA (SVDU-IPCA), to overcome the limitations of existing IPCA techniques.
  • To ensure bounded approximation errors and provide a complexity analysis for the proposed SVDU-IPCA algorithm.
  • To demonstrate the applicability and effectiveness of SVDU-IPCA in face recognition tasks.

Related Experiment Videos

Main Methods:

  • Developed a new IPCA algorithm (SVDU-IPCA) leveraging Singular Value Decomposition (SVD) updating.
  • Mathematically proved that the approximation error in the SVDU-IPCA algorithm is bounded.
  • Performed complexity analysis and evaluated the method's extensibility to a kernel version.

Main Results:

  • The proposed SVDU-IPCA method demonstrated mathematically bounded approximation errors.
  • Experimental evaluations on FERET, AR, and Yale B databases showed performance comparable to batch-mode PCA.
  • The average recognition accuracy difference between SVDU-IPCA and batch-mode PCA was less than 1%.

Conclusions:

  • The SVDU-IPCA algorithm provides a scalable and effective incremental approach to PCA for face recognition.
  • The method offers a close approximation to batch-mode PCA while mitigating computational and memory burdens.
  • SVDU-IPCA presents a promising advancement for large-scale face recognition systems.