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

770
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...
770
Convolution Properties I01:20

Convolution Properties I

508
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:
508
Convolution Properties II01:17

Convolution Properties II

537
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...
537
Neural Circuits01:25

Neural Circuits

2.5K
Neural circuits and neuronal pools are two of the main structures found in the nervous system. Neural circuits are networks of neurons that work together to carry out a specific task or process. They consist of interconnected neurons and glial cells, which provide structural and metabolic support.
Neuronal pools are collections of nerve cells with similar functions and interact through chemical and electrical signals. These pools include both interneurons (the central neural circuit nodes that...
2.5K
Deconvolution01:20

Deconvolution

511
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...
511
Accuracy and Precision01:52

Accuracy and Precision

13.8K
Scientists typically make repeated measurements of a quantity to ensure the quality of their findings and to evaluate both the precision and the accuracy of their results. Measurements are said to be precise if they yield very similar results when repeated in the same manner. A measurement is considered accurate if it yields a result that is very close to the true or the accepted value. Precise values agree with each other; accurate values agree with a true value.  Highly accurate...
13.8K

You might also read

Related Articles

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

Sort by
Same author

Restoring warped document images through 3D shape modeling.

IEEE transactions on pattern analysis and machine intelligence·2006
Same author

Lamivudine prophylaxis reduces the incidence and severity of hepatitis in hepatitis B virus carriers who receive chemotherapy for lymphoma.

Cancer·2006
Same author

Influence of an amide group in methyl octadecanoates on the monolayer stability.

Langmuir : the ACS journal of surfaces and colloids·2006
Same author

The co-existence of two growth hormone receptors in teleost fish and their differential signal transduction, tissue distribution and hormonal regulation of expression in seabream.

Journal of molecular endocrinology·2006
Same author

Proteomic profiling of regionalized proteins in rat epididymis indicates consistency between specialized distribution and protein functions.

Journal of proteome research·2006
Same author

Identification of a new Neisseria meningitidis serogroup C clone from Anhui province, China.

Lancet (London, England)·2006

Related Experiment Video

Updated: Jan 1, 2026

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

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

9.9K

IoTNet: An Efficient and Accurate Convolutional Neural Network for IoT Devices.

Tom Lawrence1, Li Zhang1

  • 1Department of Computer and Information Sciences, Faculty of Engineering and Environment, Northumbria University, Newcastle upon Tyne NE1 8ST, UK.

Sensors (Basel, Switzerland)
|December 19, 2019
PubMed
Summary

We introduce IoTNet, a novel Convolutional Neural Network (CNN) for IoT devices. IoTNet achieves state-of-the-art accuracy in efficient models by factorizing convolutions, outperforming existing models with fewer computations.

Keywords:
Convolutional Neural Networkcomputational complexitycomputer visiondeep learningdeep network architectureefficient architectureimage classification

Related Experiment Videos

Last Updated: Jan 1, 2026

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

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

9.9K

Area of Science:

  • Computer Vision
  • Deep Learning
  • Embedded Systems

Background:

  • Deploying Convolutional Neural Networks (CNNs) on resource-constrained IoT devices presents a challenge.
  • Existing approaches involve scaling down large models or using inherently small architectures that often sacrifice accuracy for computational efficiency.
  • Small architectures typically use depth-wise convolutions, trading accuracy for reduced computational cost.

Purpose of the Study:

  • To propose a novel CNN architecture, IoTNet, specifically designed for resource-constrained environments.
  • To achieve state-of-the-art performance within the domain of small, efficient models for IoT applications.
  • To introduce a new method for trading accuracy with computational cost by factorizing standard convolutions.

Main Methods:

  • IoTNet factorizes standard 3x3 convolutions into pairs of 1x3 and 3x1 convolutions, differing from depth-wise convolutions used in other efficient models.
  • Benchmarking was performed against state-of-the-art efficiency-focused models and scaled-down large architectures.
  • Performance was evaluated on datasets matching the complexity of resource-constrained problems, comparing accuracy and floating-point operations (FLOPs).

Main Results:

  • IoTNet achieved state-of-the-art accuracy improvements over MobileNetV2 on CIFAR-10 (+13.43% accuracy, -39% FLOPs).
  • Significant gains were also observed over ShuffleNet on SVHN (+6.49% accuracy, -31.8% FLOPs) and MobileNet on GTSRB (+5% accuracy, -0.38% FLOPs).
  • The results demonstrate IoTNet's superior balance of accuracy and computational efficiency.

Conclusions:

  • IoTNet offers a new, effective approach to designing efficient CNNs for resource-constrained IoT devices.
  • The convolution factorization technique provides a competitive alternative to depth-wise convolutions for achieving high accuracy with low computational cost.
  • IoTNet represents a significant advancement in efficient deep learning for edge computing applications.