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

Convolution: Math, Graphics, and Discrete Signals01:24

Convolution: Math, Graphics, and Discrete Signals

268
In any LTI (Linear Time-Invariant) system, the convolution of two signals is denoted using a convolution operator, assuming all initial conditions are zero. The convolution integral can be divided into two parts: the zero-input or natural response and the zero-state or forced response, with t0 indicating the initial time.
To simplify the convolution integral, it is assumed that both the input signal and impulse response are zero for negative time values. The graphical convolution process...
268
Convolution Properties I01:20

Convolution Properties I

155
Convolution computations can be simplified by utilizing their inherent properties.
The commutative property reveals that the input and the impulse response of an LTI (Linear Time-Invariant) system can be interchanged without affecting the output:
155
Convolution Properties II01:17

Convolution Properties II

210
The important convolution properties include width, area, differentiation, and integration properties.
The width property indicates that if the durations of input signals are T1 and T2, then the width of the output response equals the sum of both durations, irrespective of the shapes of the two functions. For instance, convolving two rectangular pulses with durations of 2 seconds and 1 second results in a function with a width of 3 seconds.
The area property asserts that the area under the...
210
Downsampling01:20

Downsampling

164
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...
164
Block Diagram Reduction01:22

Block Diagram Reduction

220
The process of deriving the transfer function of a control system often involves reducing its block diagram to a single block. This simplification can be achieved through a series of strategic operations, including relocating branch points and comparators. These operations preserve the overall function of the system while allowing for easier manipulation and combination of blocks.
The first step in this process is the identification and relocation of a branch point. A branch point, where a...
220
Upsampling01:22

Upsampling

240
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...
240

You might also read

Related Articles

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

Sort by
Same author

Frequency-Tuned Universal Adversarial Attacks on Texture Recognition.

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

Bearing Fault Diagnosis Using Lightweight and Robust One-Dimensional Convolution Neural Network in the Frequency Domain.

Sensors (Basel, Switzerland)·2022
Same author

An update on the knee osteoarthritis severity grading using wide residual learning.

Journal of X-ray science and technology·2022
Same author

It GAN Do Better: GAN-Based Detection of Objects on Images With Varying Quality.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2021
Same author

DeepCorrect: Correcting DNN Models Against Image Distortions.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2019
Same author

A Locally Weighted Fixation Density-Based Metric for Assessing the Quality of Visual Saliency Predictions.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2016
Same journal

Dynamic analysis and reliable mechanical optimization application of ring HNN effected with a memristive neuron.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

DAFF-Net: A detection and search method for small-scale low surface brightness galaxies.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Quasi-synchronization for complex networks with hybrid pinning intermittent control.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Physics-encoded convolutional neural operators for parametric PDEs: A convergence-guaranteed framework via pre-computed kernel fields.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Exploiting audio-visual modalities in videos: Object detection via multi-stage bilateral coupling network.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Reliability-aware modality completion with cross-modal distillation for federated learning with missing modalities.

Neural networks : the official journal of the International Neural Network Society·2026
See all related articles

Related Experiment Video

Updated: Jul 11, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

564

Reduced-complexity Convolutional Neural Network in the compressed domain.

Hamdan Abdellatef1, Lina J Karam2

  • 1School of Engineering - Electrical & Computer Engineering Department, Lebanese American University, Byblos, Lebanon.

Neural Networks : the Official Journal of the International Neural Network Society
|November 12, 2023
PubMed
Summary
This summary is machine-generated.

This study introduces compressed domain learning for Convolutional Neural Networks (CNNs), significantly speeding up image processing. By processing JPEG-compressed images, CNNs achieve faster speeds with comparable accuracy and reduced data storage.

Keywords:
CNNCompressed-domainDCTDeep learningFrequencyLow complexity

More Related Videos

Lensless Fluorescent Microscopy on a Chip
11:23

Lensless Fluorescent Microscopy on a Chip

Published on: August 17, 2011

17.7K
Deep Neural Networks for Image-Based Dietary Assessment
13:19

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

9.2K

Related Experiment Videos

Last Updated: Jul 11, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

564
Lensless Fluorescent Microscopy on a Chip
11:23

Lensless Fluorescent Microscopy on a Chip

Published on: August 17, 2011

17.7K
Deep Neural Networks for Image-Based Dietary Assessment
13:19

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

9.2K

Area of Science:

  • Computer Vision
  • Deep Learning
  • Image Processing

Background:

  • Deep neural networks, particularly Convolutional Neural Networks (CNNs), excel in computer vision but are computationally intensive.
  • Images are often stored and transmitted in compressed formats like JPEG, presenting challenges for traditional spatial domain CNNs.
  • Existing CNNs are slow and require significant computational resources, limiting their practical application.

Purpose of the Study:

  • To reduce the computational complexity and improve the speed of popular CNNs by performing learning and inference in the compressed domain.
  • To develop methods for efficient processing of JPEG-compressed images within neural networks.
  • To maintain classification accuracy while enhancing the efficiency of deep learning models.

Main Methods:

  • Proposed a novel graph-based frequency channel selection method to identify and retain important frequency components.
  • Implemented learning and inference directly on JPEG-compressed image data.
  • Reduced computational complexity by discarding insignificant frequency components and eliminating unnecessary network layers.
  • Introduced a preprocessing step with partial encoding to enhance resilience to compression-induced distortions.

Main Results:

  • Modified ResNet-50 operating in the compressed domain achieved up to 70% faster performance compared to spatial-domain ResNet-50.
  • Similar classification accuracy was maintained between compressed-domain and spatial-domain CNNs.
  • Training with highly compressed data resulted in good classification accuracy with up to 93% reduction in training data storage requirements.

Conclusions:

  • Performing CNN learning and inference in the compressed domain is a viable strategy to significantly improve speed and reduce computational load.
  • The proposed frequency channel selection and partial encoding methods effectively address challenges associated with compressed image processing.
  • Compressed-domain CNNs offer a promising approach for efficient and accurate computer vision tasks, especially in resource-constrained environments.