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

Local transform features and hybridization for accurate face and human detection.

Bongjin Jun1, Inho Choi, Daijin Kim

  • 1Department of Computer Science and Engineering, Pohang University of Science and Technology (POSTECH), Room 206, Engineering Bldg. II, San 31, Hyoja-dong, Nam-gu, Pohang, Gyeongbuk 790-784, Republic of Korea. simple21@postech.ac.kr

IEEE Transactions on Pattern Analysis and Machine Intelligence
|April 20, 2013
PubMed
Summary

This study introduces novel local gradient patterns (LGP) and binary histograms of oriented gradients (BHOG) features for robust object detection. A hybrid approach combining these features significantly enhances face and human detection accuracy and speed.

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

Ensemble Machine Learning- and Deep Learning-Driven Identification and Validation of Sennidin B as a Novel Dipeptidyl Peptidase-4 Inhibitor.

International journal of molecular sciences·2026
Same author

2D MoS<sub>2</sub>-conformal 3D-printed platform for dual-phototherapy and bone regeneration.

Nano convergence·2026
Same author

Therapeutic potential of Amentoflavone against myostatin for skeletal muscle atrophy treatment: An in silico, in vitro, and in vivo study.

Phytomedicine : international journal of phytotherapy and phytopharmacology·2026
Same author

Cell-free therapy for alopecia via the secretome of hiPSC-derived dermal papilla cells.

Journal of tissue engineering·2026
Same author

Optimizing chicken muscle stem cell isolation using recombinant proteolytic enzymes for cultured meat production.

Food science of animal resources·2026
Same author

Snapshot thickness profile cross-sectioning for a warped substrate with non-uniform sub-100 nm ultra-thin film.

Optics express·2026

Area of Science:

  • Computer Vision
  • Machine Learning
  • Pattern Recognition

Background:

  • Accurate and efficient detection of faces and humans is crucial in various applications.
  • Existing methods often struggle with variations in illumination and pose.
  • Local transform features offer robustness to such variations but require efficient computation.

Purpose of the Study:

  • To propose novel local transform features: Local Gradient Patterns (LGP) and Binary Histograms of Oriented Gradients (BHOG).
  • To develop a hybrid feature combining LGP, BHOG, and other local features using AdaBoost.
  • To evaluate the performance of these features for face and human detection.

Main Methods:

  • Local Gradient Patterns (LGP) feature: Assigns a binary value based on local gradient comparison to its neighbors.

Related Experiment Videos

  • Binary Histograms of Oriented Gradients (BHOG) feature: Uses a binary thresholding on histogram bins for efficiency.
  • Hybrid feature: Integrates LGP, BHOG, and other local features via AdaBoost for sequential selection of optimal discriminative features.

    • Main Results:

    • LGP demonstrates robustness to local intensity variations.
    • BHOG achieves fast computation times without post-processing.
    • The hybrid feature significantly improves face and human detection accuracy, robust to global illumination and local pose changes.
    • Experiments on MIT+CMU, FDDB, INRIA, and Caltech databases confirm the effectiveness of the proposed features.

    Conclusions:

    • The proposed LGP and BHOG features provide accurate detection and efficient computation, respectively.
    • The hybrid feature effectively combines the strengths of individual local transform features.
    • This approach offers a considerable improvement in face and human detection performance, particularly under challenging conditions.