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

Scalable Face Image Retrieval with Identity-Based Quantization and Multireference Reranking.

Zhong Wu, Qifa Ke, Jian Sun

    IEEE Transactions on Pattern Analysis and Machine Intelligence
    |June 8, 2011
    PubMed
    Summary
    This summary is machine-generated.

    Related Concept Videos

    Introduction to Scalers01:21

    Introduction to Scalers

    Many familiar physical quantities can be specified completely by giving a single number and the appropriate unit. For example, "a class period lasts 50 min," or "the gas tank in my car holds 65 L," or "the distance between the two posts is 100 m." A physical quantity that can be specified completely in this manner is called a scalar quantity. The word "scalar" is a synonym for "number." Time, mass, distance, length, volume, temperature, and energy are some examples of scalar quantities.
    Scalar...

    You might also read

    Related Articles

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

    Sort by
    Same author

    CTRL: Closed-Loop Transcription to an LDR via Minimaxing Rate Reduction.

    Entropy (Basel, Switzerland)·2022
    Same author

    Double-barrel vascularised fibula graft in mandibular reconstruction: a 10-year experience with an algorithm.

    Journal of plastic, reconstructive & aesthetic surgery : JPRAS·2012
    Same author

    Expression of ERBB3 binding protein 1 (EBP1) in salivary adenoid cystic carcinoma and its clinicopathological relevance.

    BMC cancer·2012
    Same author

    Introgression and selection shaping the genome and adaptive loci of weedy rice in northern China.

    The New phytologist·2012
    Same author

    Intramolecular Schmidt reaction of acyl chlorides with alkyl azides: rapid access to fused polycyclic nitrogen-containing heterocycles via a multistep one-pot transformation.

    Organic letters·2012
    Same author

    Prohibitin induces apoptosis in BGC823 gastric cancer cells through the mitochondrial pathway.

    Asian Pacific journal of cancer prevention : APJCP·2012
    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

    This study introduces a scalable face image retrieval system using novel local and global features. The new method improves retrieval accuracy and efficiency for large face databases.

    Area of Science:

    • Computer Science
    • Biometrics
    • Image Processing

    Background:

    • Traditional image retrieval struggles with face-specific data due to low-discriminative visual words and ignored facial properties.
    • High-dimensional face recognition features are unsuitable for scalable indexing, leading to computational and storage limitations.

    Purpose of the Study:

    • To develop a scalable face image retrieval system that overcomes the limitations of existing methods.
    • To create a novel face representation combining local and global features for improved retrieval performance.

    Main Methods:

    • Developed component-based local features exploiting face properties and quantized them into visual words using identity-based quantization.
    • Encoded discriminative global features using a compact 40-byte Hamming signature for each face.

    Related Experiment Videos

  • Implemented a retrieval stage using an inverted index for initial candidate retrieval, followed by multi-reference distance reranking with Hamming signatures.

    • Main Results:

    • Demonstrated that local features provide good recall via inverted indexing, while global Hamming signatures enhance precision through reranking.
    • Showcased complementary performance of local and global features on a one-million face database.
    • Achieved superior retrieval quality compared to linear scan methods using state-of-the-art face recognition features.

    Conclusions:

    • The proposed system is highly scalable for large-scale face image retrieval.
    • The combination of local and global features significantly improves both recall and precision in face retrieval.
    • This approach offers a more effective and efficient solution for face image retrieval tasks.