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

Shape from defocus via diffusion.

Paolo Favaro1, Stefano Soatto, Martin Burger

  • 1School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh, UK. p.favaro@hw.ac.uk

IEEE Transactions on Pattern Analysis and Machine Intelligence
|January 16, 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

Acute Upper Gastrointestinal Bleeding is Associated With Poor Prognosis in Patients With Biliary Tract Cancer.

Cancer medicine·2026
Same author

AI Agents as Universal Task Solvers.

Entropy (Basel, Switzerland)·2026
Same author

Acute Variceal Bleeding During the SARS-CoV-2 Pandemic: A National Multicenter Observational Study.

Journal of clinical medicine·2025
Same author

Analysis of mean-field models arising from self-attention dynamics in transformer architectures with layer normalization.

Philosophical transactions. Series A, Mathematical, physical, and engineering sciences·2025
Same author

Quality in science communication with communicative artificial intelligence: A principle-based framework.

Public understanding of science (Bristol, England)·2025
Same author

Fine-tune language models as multi-modal differential equation solvers.

Neural networks : the official journal of the International Neural Network Society·2025
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 models image defocus as a heat diffusion process. By analyzing relative blur between two images, researchers developed a novel method to estimate 3-D scene depth without deblurring.

Area of Science:

  • Computer Vision
  • Image Processing
  • Computational Imaging

Background:

  • Image defocus and blur can be mathematically modeled using the heat equation, analogous to heat diffusion.
  • Extending this to non-planar scenes requires a space-varying diffusion coefficient.
  • Reconstructing 3-D structure from blurred images (inverse problem) is ill-posed.

Purpose of the Study:

  • To develop a novel method for estimating 3-D scene depth from blurred images.
  • To bypass the ill-posed nature of inverse diffusion problems in 3-D reconstruction.
  • To utilize the concept of relative blur for depth estimation.

Main Methods:

  • Modeling defocus as a diffusion process using the heat equation.
  • Extending the diffusion model to space-varying diffusion coefficients for non-planar scenes.

Related Experiment Videos

  • Employing the notion of relative blur between two images to infer depth.

    • Main Results:

    • Demonstrated that the amount of diffusion needed to blur an image is related to scene depth.
    • Developed a global algorithm for depth profile estimation.
    • Successfully estimated depth without explicitly recovering deblurred images, using only forward diffusion.

    Conclusions:

    • Relative blur provides a viable approach to circumvent ill-posed inverse diffusion problems in 3-D scene reconstruction.
    • The proposed method enables direct depth estimation from blurred image pairs.
    • This technique offers a new pathway for 3-D scene understanding in computer vision.