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 geometric approach to shape from defocus.

P Favaro, S Soatto

    IEEE Transactions on Pattern Analysis and Machine Intelligence
    |March 8, 2005
    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

    3-D retinal surface inference: stereo or monocular fundus camera?

    Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference·2007
    Same author

    [Pharmacoeconomic aspects of burn therapies].

    Annales pharmaceutiques francaises·2002
    Same author

    [Hormone replacement therapy and the wellbeing of menopausal women].

    Minerva ginecologica·1999
    Same author

    [Relation between Doppler echo-flowmetry and oncogenic markers in pelvic masses].

    Minerva ginecologica·1999
    Same author

    Increasing the relevancy and usefulness of programme evaluation: a comprehensive evaluation.

    Hygie·1988
    Same author

    Individual differences in voluntary heart rate control.

    Archiv fur Psychologie·1984
    Same journal

    TraGraph-GS: Trajectory Graph-based Gaussian Splatting for Arbitrary Large-Scale Scene Rendering.

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

    SWIFT: A Small-World Interaction Framework for Flow-Aware Trajectory Prediction in Autonomous Driving.

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

    This study presents a new method for 3D shape from defocus, estimating scene geometry from blurred images without needing to deblur them or make restrictive assumptions. The approach is efficient and robust, performing well even when the point spread function is unknown.

    Area of Science:

    • Computer Vision
    • Computational Imaging
    • 3D Reconstruction

    Background:

    • Shape from defocus infers 3D scene geometry from blurred images.
    • Existing methods often require deblurring or rely on the equifocal assumption.
    • Accurate modeling of the point spread function (PSF) is crucial.

    Purpose of the Study:

    • To develop a novel approach for shape from defocus that bypasses the need for deblurring.
    • To estimate 3D geometry without strong assumptions about the scene.
    • To provide efficient and robust algorithms for 3D reconstruction from defocused images.

    Main Methods:

    • Developed an optimal method using orthogonal operators regularized by functional singular value decomposition when the PSF is known.
    • Proposed a simple and efficient method for unknown PSF: learning projection operators from blurred images.

    Related Experiment Videos

  • Minimized the Euclidean norm of image differences via projections onto linear subspaces.

    • Main Results:

    • The proposed methods successfully estimate 3D geometry from defocused images without deblurring.
    • Performance is robust and relatively insensitive to the specific form of the point spread function.
    • Algorithms involve simple matrix-vector multiplications, enabling real-time implementation.

    Conclusions:

    • The novel shape from defocus approach eliminates the necessity of deblurring and restrictive scene assumptions.
    • The methods are geometrically grounded, utilizing Hilbert space structures for efficient computation.
    • The algorithms offer a practical and real-time solution for 3D reconstruction from defocused imagery.