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

Dynamic directional gradient vector flow for snakes.

Jierong Cheng1, Say Wei Foo

  • 1School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore. jpcheng@pmail.ntu.edu.sg

IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
|June 13, 2006
PubMed
Summary

This study introduces a new dynamic directional gradient vector flow (DDGVF) snake model for improved image segmentation. The DDGVF snake effectively distinguishes edge types, outperforming traditional methods in complex scenarios.

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

High-efficiency terahertz beam scanning based on topologically optimized freestanding metasurfaces.

Optics letters·2025
Same author

Terahertz Multi-State Reconfigurable Holographic Metasurface Based on Full-Stokes Magnetic Eigen-Modulation of Photonic Weyl Semimetal.

Advanced materials (Deerfield Beach, Fla.)·2025
Same author

Robust terahertz polarization analysis based on a unit-scale polarization-multiplexed metasurface.

Optics letters·2025
Same author

Broadband large-angle beam scanning with dynamic spin energy distribution based on liquid crystal cascaded bilayer metasurface.

Nanophotonics (Berlin, Germany)·2024
Same author

Flexible terahertz phase shifter for optically controlled polydimethylsiloxane-vanadium dioxide composite film.

Optics express·2024
Same author

Flexible broadband metamaterial absorber in long-wave infrared with visible transparency fabricated by laser direct writing.

Optics letters·2023

Area of Science:

  • Computer Vision
  • Image Processing
  • Computational Geometry

Background:

  • Active contour models, or snakes, are prevalent in image segmentation.
  • Current snake models struggle to differentiate between positive and negative step edges, limiting segmentation accuracy.
  • Gradient Vector Flow (GVF) snakes offer improvements but still face challenges with complex edge orientations.

Purpose of the Study:

  • To propose a novel dynamic external force for snakes, termed dynamic directional gradient vector flow (DDGVF).
  • To enhance image segmentation performance by enabling snakes to discern between positive and negative step edges.
  • To improve snake model robustness in segmenting images with edges of varying directions.

Main Methods:

  • Development of the dynamic directional gradient vector flow (DDGVF) external force.

Related Experiment Videos

  • Separate handling of external force fields for x and y gradient directions.
  • Dynamic utilization of the DDGVF field based on snake orientation during deformation iterations.

    • Main Results:

    • The DDGVF snake model demonstrates superior segmentation compared to the standard GVF snake.
    • Improved performance is observed in segmenting images containing edges with diverse orientations.
    • The model successfully addresses segmentation confusion caused by complex edge directional variations.

    Conclusions:

    • The proposed DDGVF snake model offers a significant advancement in image segmentation accuracy.
    • This method effectively overcomes limitations of existing snake models in handling varied edge directions.
    • DDGVF provides a more robust and precise tool for image segmentation applications.