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

Training an active random field for real-time image denoising.

Adrian Barbu1

  • 1Department of Statistics, Florida State University, Tallahassee, FL 32306, USA. abarbu@stat.fsu.edu

IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
|July 29, 2009
PubMed
Summary

Training Markov random field (MRF) or conditional random field (CRF) models with fast inference algorithms, termed active random fields (ARF), significantly boosts computer vision task speed and accuracy. This approach enables real-time image denoising with state-of-the-art performance.

Related Concept Videos

Deconvolution01:20

Deconvolution

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...

You might also read

Related Articles

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

Sort by
Same author

MSLNet and Perceptual Grouping for Guidewire Segmentation and Localization.

Sensors (Basel, Switzerland)·2025
Same author

Training a Two-Layer ReLU Network Analytically.

Sensors (Basel, Switzerland)·2023
Same author

The Compact Support Neural Network.

Sensors (Basel, Switzerland)·2021
Same author

Are screening methods useful in feature selection? An empirical study.

PloS one·2019
Same author

Prolonged treatment with open-label pirfenidone in Hermansky-Pudlak syndrome pulmonary fibrosis.

Molecular genetics and metabolism·2018
Same author

Feature Selection with Annealing for Computer Vision and Big Data Learning.

IEEE transactions on pattern analysis and machine intelligence·2016

Area of Science:

  • Computer Vision
  • Machine Learning
  • Bayesian Inference

Background:

  • Markov random fields (MRF) and conditional random fields (CRF) are common Bayesian frameworks for computer vision.
  • Typically, MRF/CRF models are trained separately from their inference algorithms.
  • This separation can lead to suboptimal performance and slow processing speeds.

Purpose of the Study:

  • To investigate the benefits of jointly training MRF/CRF models with fast inference algorithms.
  • To introduce the concept of Active Random Fields (ARF) for integrated model and inference training.
  • To apply the ARF approach to image denoising and evaluate its performance.

Main Methods:

  • Defined Active Random Fields (ARF) as a unified model combining MRF/CRF with a fast inference algorithm.

Related Experiment Videos

  • Trained the ARF using a loss function optimization on a dataset of input images and desired outputs.
  • Applied ARF to image denoising using the Fields of Experts MRF and a gradient descent inference algorithm (1-4 iterations).

    • Main Results:

    • Achieved significant gains in both speed and accuracy compared to traditional independent training.
    • Demonstrated a 1000-3000 times speedup in image denoising using the ARF approach.
    • Real-time image denoising (8fps on a single CPU for 256x256 images) with near state-of-the-art accuracy was achieved.

    Conclusions:

    • Jointly training MRF/CRF models with fast inference algorithms (ARF) offers substantial performance improvements.
    • The ARF approach enables efficient and accurate image denoising, suitable for real-time applications.
    • This integrated training strategy represents a promising direction for advancing computer vision techniques.