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

Cascaded linear shift-invariant processors in optical pattern recognition.

S Reed, J Coupland

    Applied Optics
    |March 25, 2008
    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

    The global extent of the grassland biome and implications for the terrestrial carbon sink.

    Nature ecology & evolution·2026
    Same author

    Serotonergic psychedelics for depression: A comprehensive overview.

    International review of neurobiology·2025
    Same author

    Prevalence of Sickle Cell Disease among Newborns in St Vincent and the Grenadines: A Retrospective Study.

    The West Indian medical journal·2024
    Same author

    5 Maternal gestational nutrition perturbs small RNA code in offspring sperm in sheep.

    Reproduction, fertility, and development·2022
    Same author

    Benign adenoidal hypertrophy caused by adenovirus presenting as a nasopharyngeal mass concerning for malignancy.

    International journal of pediatric otorhinolaryngology·2020
    Same author

    Acute and chronic effects of multivitamin/mineral supplementation on objective and subjective energy measures.

    Nutrition & metabolism·2020
    Same journal

    Multifunctional reconfigurable terahertz metasurface based on vanadium dioxide phase transition: achieving broadband absorption and efficient polarization conversion.

    Applied optics·2026
    Same journal

    High-Q-factor electromagnetically induced transparency utilizing quasi-bound states in the continuum in an all-dielectric terahertz metasurface.

    Applied optics·2026
    Same journal

    Automated stitching interferometry for high-precision metrology of X-ray mirrors.

    Applied optics·2026
    Same journal

    Experimental demonstration of an approach to designing a metal-dielectric DBR resonant cavity structure.

    Applied optics·2026
    Same journal

    High-precision wavefront reconstruction from a single-shot interferogram using a physics-driven hybrid feature calibration network.

    Applied optics·2026
    Same journal

    Ultra-high-Q Fano resonance based on coupled topological corner states in Kagome photonic crystals.

    Applied optics·2026
    See all related articles

    This study introduces a novel cascaded architecture for optical pattern recognition, enhancing performance with nonlinear thresholds. This method significantly outperforms standard linear filters in defect detection tasks.

    Area of Science:

    • Optical pattern recognition
    • Neural network architectures
    • Image processing

    Background:

    • Cascaded correlators with nonlinear thresholds offer improved high-speed optical pattern recognition.
    • This architecture is a specialized form of multilayer, feed-forward neural networks.
    • The importance of this specific cascaded configuration for shift-invariant outputs has been underestimated.

    Purpose of the Study:

    • To investigate and validate the effectiveness of a cascaded architecture of linear shift-invariant processing modules with nonlinear thresholds.
    • To demonstrate that this cascaded structure is optimal for multilayer feed-forward neural networks requiring shift-invariant outputs.
    • To compare the performance of this nonlinear cascade against traditional linear filters in a practical application.

    Main Methods:

    Related Experiment Videos

    • Development and digital modeling of a cascaded system using linear correlators and nonlinear thresholds.
    • Training the digital model with a modified backpropagation algorithm and simulated annealing for optimization.
    • Benchmarking the system's performance on a defect recognition task in the canning industry.

    Main Results:

    • The cascaded architecture was proven to be the exact structure for shift-invariant outputs in multilayer feed-forward neural networks.
    • The proposed system is practical for optical implementation, suitable for high-speed analysis of large images.
    • The nonlinear cascade demonstrated significant performance improvement over the standard linear minimum average correlation energy (MACE) filter in defect recognition.

    Conclusions:

    • The cascaded nonlinear architecture provides a significant performance advantage over linear methods for optical pattern recognition.
    • This approach is well-suited for high-speed image analysis due to its optical implementability.
    • The study highlights the overlooked potential of cascaded correlators for advanced pattern recognition tasks.