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

Tensor-based AAM with continuous variation estimation: application to variation-robust face recognition.

Hyung-Soo Lee1, Daijin Kim

  • 1Research Lab., Olaworks, Inc., 738-1, Yeoksam 1-dong, Gangnam-gu, Seoul 135-924, Korea. sooz@olaworks.com

IEEE Transactions on Pattern Analysis and Machine Intelligence
|April 18, 2009
PubMed
Summary
This summary is machine-generated.

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

Disability and Relapse Risk in Late-Onset Myelin Oligodendrocyte Glycoprotein Antibody-Associated Disease.

JAMA network open·2026
Same author

Outcomes of immunosuppressive therapy discontinuation in patients with myelin oligodendrocyte glycoprotein antibody-associated disease.

Multiple sclerosis (Houndmills, Basingstoke, England)·2025
Same author

Impact of COVID-19 Infection and Its Association With Previous Vaccination in Patients With Myasthenia Gravis in Korea: A Multicenter Retrospective Study.

Journal of Korean medical science·2024
Same author

Serial Nerve Conduction Studies in Guillain-Barré Syndrome: Its Usefulness and Precise Timing.

Journal of clinical neurophysiology : official publication of the American Electroencephalographic Society·2024
Same author

More than autophony: a case of Kennedy's disease presenting with autophony as an early clinical manifestation.

The Journal of laryngology and otology·2023
Same author

Investigating the Precise Identification of Citrullination Sites with High- Performance Score Metrics Using a Powerful Computation Predicting Tool.

Combinatorial chemistry & high throughput screening·2023
Same journal

HardFlow: Hard-Constrained Sampling for Flow-Matching Models Via Trajectory Optimization.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Industrial Brain: Self-Evolving Neuro-Symbolic Autonomy with Causal Resilience for Cyber-Physical Systems.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Adaptive Hardness-Driven Dictionary Distillation for Incomplete Streaming View Clustering.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Mixture of Global and Local Experts with Diffusion Transformer for Controllable Face Generation.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Task-KV: Task-aware KV Cache Optimization via Semantic Differentiation of Attention Heads.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Achieving Text-based Person Retrieval with Any Granularity.

IEEE transactions on pattern analysis and machine intelligence·2026
See all related articles

A new tensor-based Active Appearance Model (AAM) improves fitting accuracy for non-rigid objects. This robust model enhances face recognition by effectively handling variations in pose, expression, and illumination.

Area of Science:

  • Computer Vision
  • Machine Learning
  • Image Processing

Background:

  • Active Appearance Models (AAMs) are effective for non-rigid object representation.
  • Conventional AAMs struggle with fitting accuracy when input images deviate significantly from training data due to fixed models.
  • Variations in pose, expression, and illumination commonly degrade AAM fitting performance.

Purpose of the Study:

  • To develop a more robust Active Appearance Model (AAM) capable of handling significant image variations.
  • To introduce a tensor-based framework for AAM fitting to improve accuracy and robustness.
  • To evaluate the performance of the proposed tensor-based AAM in variation-robust face recognition.

Main Methods:

  • Proposed a tensor-based Active Appearance Model (AAM) using an image tensor and a model tensor.

Related Experiment Videos

  • Implemented discrete and continuous variation estimation techniques within the image tensor.
  • Generated variation-specific AAM basis vectors using the model tensor.
  • Employed indirect AAM feature transformation for face recognition tasks.

    • Main Results:

    • The tensor-based AAM demonstrated improved fitting accuracy compared to conventional AAMs.
    • Continuous variation estimation within the tensor-based AAM yielded superior results over discrete estimation.
    • The proposed method achieved higher face recognition rates, demonstrating its variation-robustness.

    Conclusions:

    • The tensor-based Active Appearance Model (AAM) offers a significant improvement in fitting robustness and accuracy.
    • Continuous variation estimation is a more effective approach for handling image variations in AAMs.
    • The developed tensor-based AAM is highly effective for variation-robust face recognition applications.