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

Learning shape from shading by a multilayer network.

G Q Wei1, G Hirzinger

  • 1Inst. of Robotics and Syst. Dynamics, German Aerosp. Res. Establ., Oberpfaffenhofen.

IEEE Transactions on Neural Networks
|January 1, 1996
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

[Safety and efficacy of humanized CD19-targeted CAR-T cells in patients with relapsed/refractory acute B cell lymphoblastic leukemia].

Zhonghua xue ye xue za zhi = Zhonghua xueyexue zazhi·2023
Same author

[Jinghuaweikang capsules combined with Quadruple therapy in the treatment of <i>Helicobacter pylori</i> infection:a multicenter,randomized,controlled,clinical study].

Zhonghua yi xue za zhi·2019
Same author

[The effect of WT1 expression on the prognosis of allogeneic hematopoietic stem cell transplantation in acute leukemia].

Zhonghua xue ye xue za zhi = Zhonghua xueyexue zazhi·2019
Same author

[Application of scar skin dilatation in repair of auricular defects after burn].

Lin chuang er bi yan hou tou jing wai ke za zhi = Journal of clinical otorhinolaryngology head and neck surgery·2018
Same author

Characterization and functional analysis of serpin-1 like gene from oak silkworm Antheraea pernyi.

Bulletin of entomological research·2017
Same author

Molecular characterization, tissue distribution, and immune reaction expression of karyopherins in the domestic silkworm (Bombyx mori).

Genetics and molecular research : GMR·2015
Same journal

Universal perceptron and DNA-like learning algorithm for binary neural networks: LSBF and PBF implementations.

IEEE transactions on neural networks·2013
Same journal

Guest editorial: special section on white box nonlinear prediction models.

IEEE transactions on neural networks·2011
Same journal

Data-based fault-tolerant control of high-speed trains with traction/braking notch nonlinearities and actuator failures.

IEEE transactions on neural networks·2011
Same journal

Guest editorial: special section on data-based control, modeling, and optimization.

IEEE transactions on neural networks·2011
Same journal

Neural network-based multiple robot simultaneous localization and mapping.

IEEE transactions on neural networks·2011
Same journal

Data-driven model-free adaptive control for a class of MIMO nonlinear discrete-time systems.

IEEE transactions on neural networks·2011
See all related articles

This study introduces a novel method for shape from shading using multilayer feedforward networks without explicit example data. The approach reformulates surface reconstruction as an optimization problem, enabling accurate depth estimation and light source direction recovery.

Area of Science:

  • Computer Vision
  • Machine Learning
  • Artificial Intelligence

Background:

  • Multilayer feedforward networks typically require explicit input-output data for learning nonlinear mappings.
  • The shape from shading problem in computer vision aims to infer surface depth from image intensity variations.

Purpose of the Study:

  • To present a novel application of multilayer feedforward networks for shape from shading without explicit training data.
  • To reformulate the shape from shading problem as a network weight optimization task.

Main Methods:

  • Utilizing multilayer feedforward networks as a parametric surface representation.
  • Minimizing an error function over network weights using stochastic gradient and conjugate gradient methods.
  • Imposing boundary conditions on surface depth or normal by adjusting network levels.

Related Experiment Videos

Main Results:

  • Successfully estimated surface depth and light source direction from differential equations.
  • Demonstrated the method's efficiency with both synthetic and real image datasets.
  • Showcased the ability to integrate surface and light source estimation.

Conclusions:

  • The proposed method offers an effective way to solve shape from shading using neural networks without explicit examples.
  • The approach is versatile and can be extended to a broader range of computer vision problems.
  • Validates the practical applicability and efficiency of the developed technique.