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 Concept Videos

Upsampling01:22

Upsampling

743
Managing signal sampling rates is essential in digital signal processing to maintain signal integrity. A decimated signal, characterized by a reduced frequency range due to its lower sampling rate, can be upsampled by inserting zeros between each sample. This upsampling process expands the original spectrum and introduces repeated spectral replicas at intervals dictated by the new Nyquist frequency. To refine this zero-inserted sequence, it is passed through a lowpass filter with a cutoff...
743
Deconvolution01:20

Deconvolution

764
Deconvolution, also known as inverse filtering, is the process of extracting the impulse response from known input and output signals. This technique is vital in scenarios where the system's characteristics are unknown, and they must be inferred from the observable signals.
Deconvolution involves several mathematical techniques to derive the impulse response. One common approach is polynomial division. In this method, the input and output sequences are treated as coefficients of...
764
Downsampling01:20

Downsampling

868
When considering a sampled sequence with zero values between sampling instants, one can replace it by taking every N-th value of the sequence. At these integer multiples of N, the original and sampled sequences coincide. This process, known as decimation, involves extracting every N-th sample from a sequence, thereby creating a more efficient sequence.
The Fourier transform of the decimated sequence reveals a combination of scaled and shifted versions of the original spectrum. This...
868

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Asymptotic theory of in-context learning by linear attention.

Proceedings of the National Academy of Sciences of the United States of America·2025
Same author

Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance Fields.

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

DROID: Driver-Centric Risk Object Identification.

IEEE transactions on pattern analysis and machine intelligence·2023
Same author

Graph-Based Depth Denoising & Dequantization for Point Cloud Enhancement.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2022
Same author

Unique Geometry and Texture From Corresponding Image Patches.

IEEE transactions on pattern analysis and machine intelligence·2021
Same author

Phase Transitions in Transfer Learning for High-Dimensional Perceptrons.

Entropy (Basel, Switzerland)·2021

Related Experiment Video

Updated: Apr 25, 2026

Whole-cell Super-Resolution Imaging via DNA-PAINT on a Spinning Disk Confocal with Optical Photon Reassignment
07:12

Whole-cell Super-Resolution Imaging via DNA-PAINT on a Spinning Disk Confocal with Optical Photon Reassignment

Published on: January 6, 2026

755

Monte Carlo non-local means: random sampling for large-scale image filtering.

Stanley H Chan, Todd Zickler, Yue M Lu

    IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
    |August 15, 2014
    PubMed
    Summary

    We introduce Monte Carlo nonlocal means (MCNLM), a faster randomized algorithm for image filtering. MCNLM significantly reduces runtime for large-scale image processing while maintaining high accuracy, making it competitive with existing methods.

    More Related Videos

    Super-resolution Imaging of Neuronal Dense-core Vesicles
    09:30

    Super-resolution Imaging of Neuronal Dense-core Vesicles

    Published on: July 2, 2014

    9.2K

    Related Experiment Videos

    Last Updated: Apr 25, 2026

    Whole-cell Super-Resolution Imaging via DNA-PAINT on a Spinning Disk Confocal with Optical Photon Reassignment
    07:12

    Whole-cell Super-Resolution Imaging via DNA-PAINT on a Spinning Disk Confocal with Optical Photon Reassignment

    Published on: January 6, 2026

    755
    Super-resolution Imaging of Neuronal Dense-core Vesicles
    09:30

    Super-resolution Imaging of Neuronal Dense-core Vesicles

    Published on: July 2, 2014

    9.2K

    Area of Science:

    • Computer Vision
    • Image Processing
    • Machine Learning

    Background:

    • Classical nonlocal means (NLM) algorithm is effective for image filtering but computationally expensive for large datasets.
    • Large-scale image processing and denoising tasks require efficient algorithms to reduce computational burden.

    Purpose of the Study:

    • To develop a faster, randomized version of the NLM algorithm for large-scale image filtering.
    • To analyze the performance and accuracy of the proposed Monte Carlo nonlocal means (MCNLM) algorithm.
    • To derive optimal sampling strategies for MCNLM to minimize error probability.

    Main Methods:

    • Proposed the Monte Carlo nonlocal means (MCNLM) algorithm, a randomized approach to NLM.
    • Analyzed MCNLM performance, providing error probability bounds that decay exponentially with image/database size.
    • Derived formulas for optimal sampling patterns to minimize error probability, leveraging pairwise similarity weights.

    Main Results:

    • MCNLM demonstrates tight concentration of random outcomes around the deterministic NLM result for large images/databases.
    • Error probability bounds show exponential decay, ensuring high accuracy even with reduced sampling.
    • Numerical experiments confirm MCNLM's competitiveness with state-of-the-art fast NLM algorithms for denoising.
    • MCNLM achieved near-optimal denoising results (within 0.2 dB of full NLM) with a three-order-of-magnitude runtime reduction on a massive external database.

    Conclusions:

    • MCNLM offers a significant speedup for large-scale image filtering compared to classical NLM.
    • The algorithm provides a robust trade-off between computational efficiency and solution accuracy.
    • MCNLM is a promising alternative for real-time and large-scale image denoising applications.