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

Performance evaluation of local descriptors.

Krystian Mikolajczyk1, Cordelia Schmid

  • 1Department of Engineering Science, University of Oxford, Oxford, OX1 3PJ, United Kingdom. km@robots.ox.ac.uk

IEEE Transactions on Pattern Analysis and Machine Intelligence
|October 22, 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

Match Stereo Videos Via Bidirectional Alignment.

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

CoVR-2: Automatic Data Construction for Composed Video Retrieval.

IEEE transactions on pattern analysis and machine intelligence·2024
Same author

Learning to Answer Visual Questions From Web Videos.

IEEE transactions on pattern analysis and machine intelligence·2022
Same author

Key.Net: Keypoint Detection by Handcrafted and Learned CNN Filters Revisited.

IEEE transactions on pattern analysis and machine intelligence·2022
Same author

On the Importance of Visual Context for Data Augmentation in Scene Understanding.

IEEE transactions on pattern analysis and machine intelligence·2019
Same author

[Formula: see text]-Patches: A Benchmark and Evaluation of Handcrafted and Learned Local Descriptors.

IEEE transactions on pattern analysis and machine intelligence·2019
Same journal

Benchmarking the Robustness of Autonomous Driving to Environmental Illusions: A Lane Perception Perspective.

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

Learning Topology-Aware Representations via Test-Time Adaptation for Anomaly Segmentation.

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

This study compares image descriptors for local interest regions, finding that SIFT-based descriptors offer the best performance. The research also introduces an improved SIFT descriptor that surpasses the original method.

Area of Science:

  • Computer Vision
  • Image Processing
  • Machine Learning

Background:

  • Numerous local image descriptors exist, but their comparative performance and dependence on interest region detectors are unclear.
  • Key requirements for descriptors include distinctiveness and robustness to viewing condition variations and detector errors.

Purpose of the Study:

  • To evaluate and compare the performance of various local image descriptors.
  • To determine the influence of different interest region detectors on descriptor performance.
  • To propose and validate an enhanced SIFT descriptor.

Main Methods:

  • Performance evaluation based on recall with respect to precision under various image transformations.
  • Comparison of descriptors including Shape Context, Steerable Filters, PCA-SIFT, Differential Invariants, Spin Images, SIFT, Complex Filters, Moment Invariants, and Cross-correlation.

Related Experiment Videos

  • Assessment across different types of interest regions, such as those detected by Harris-Affine.

    • Main Results:

    • SIFT-based descriptors demonstrated the highest performance across evaluations.
    • An extended SIFT descriptor was proposed and showed superior performance compared to the original SIFT.
    • The ranking of descriptor performance was largely independent of the chosen interest region detector.
    • Moments and steerable filters performed best among low-dimensional descriptors.

    Conclusions:

    • SIFT-based descriptors are highly effective for local interest region description.
    • The proposed SIFT extension offers improved distinctiveness and robustness.
    • Descriptor performance ranking is relatively stable across different interest region detectors.