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

Improving Translational Accuracy02:07

Improving Translational Accuracy

10.4K
Base complementarity between the three base pairs of mRNA codon and the tRNA anticodon is not a failsafe mechanism. Inaccuracies can range from a single mismatch to no correct base pairing at all. The free energy difference between the correct and nearly correct base pairs can be as small as 3 kcal/ mol. With complementarity being the only proofreading step, the estimated error frequency would be one wrong amino acid in every 100 amino acids incorporated. However, error frequencies observed in...
10.4K
Force Classification01:22

Force Classification

1.2K
Forces play a crucial role in the study of physics and engineering. They are essential in describing the motion, behavior, and equilibrium of objects in the physical world. Forces can be classified based on their origin, type, and direction of action.
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...
1.2K
Aggregates Classification01:29

Aggregates Classification

321
Aggregate classification is generally based on its size, petrographic characteristics, weight, and source. Size classification ranges from coarse to fine aggregates, defined by the size of the particles. Coarse aggregates are particles that do not pass through ASTM sieve No. 4, and aggregates that pass through the sieve are fine aggregates.
Petrographic classification groups aggregates based on common mineralogical characteristics. Some of the common mineral groups found in aggregates are...
321
Classification of Signals01:30

Classification of Signals

460
In signal processing, signals are classified based on various characteristics: continuous-time versus discrete-time, periodic versus aperiodic, analog versus digital, and causal versus noncausal. Each category highlights distinct properties crucial for understanding and manipulating signals.
A continuous-time signal holds a value at every instant in time, representing information seamlessly. In contrast, a discrete-time signal holds values only at specific moments, often denoted as x(n), where...
460
Classification of Systems-I01:26

Classification of Systems-I

186
Linearity is a system property characterized by a direct input-output relationship, combining homogeneity and additivity.
Homogeneity dictates that if an input x(t) is multiplied by a constant c, the output y(t) is multiplied by the same constant. Mathematically, this is expressed as:
186
Classification of Systems-II01:31

Classification of Systems-II

146
Continuous-time systems have continuous input and output signals, with time measured continuously. These systems are generally defined by differential or algebraic equations. For instance, in an RC circuit, the relationship between input and output voltage is expressed through a differential equation derived from Ohm's law and the capacitor relation,
146

You might also read

Related Articles

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

Sort by
Same author

Enhancing reproductive health clinical training through integrated pedagogy: A quasi-experimental study.

African journal of reproductive health·2025
Same author

Multimodal fusion image enhancement technique and CFEC-YOLOv7 for underwater target detection algorithm research.

Frontiers in neurorobotics·2025
Same author

<i>O</i>-Cyanobenzaldehydes Irreversibly Modify Both Buried and Exposed Lysine Residues in Live Cells.

Journal of the American Chemical Society·2025
Same author

Discovery of 5-aminopyrido[2,3-d]pyrimidin-7(8H)-one derivatives as new hematopoietic progenitor kinase 1 (HPK1) inhibitors.

European journal of medicinal chemistry·2024
Same author

Pooling region learning of visual word for image classification using bag-of-visual-words model.

PloS one·2020
Same author

Protective effect of alpha-mangostin against oxidative stress induced-retinal cell death.

Scientific reports·2016

Related Experiment Video

Updated: Jul 2, 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

533

Can using a pre-trained deep learning model as the feature extractor in the bag-of-deep-visual-words model always

Ye Xu1, Xin Zhang1, Chongpeng Huang1

  • 1School of IoT Technology, Wuxi Institute of Technology, Wuxi, Jiangsu, China.

Plos One
|February 29, 2024
PubMed
Summary

Using pre-trained deep learning models as feature extractors in Bag-of-Deep-Visual-Words models can improve classification accuracy, especially with Fisher Vector encoding. This approach shows significant potential for enhancing performance across various datasets and models.

More Related Videos

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
Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images
08:20

Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images

Published on: October 27, 2023

1.4K

Related Experiment Videos

Last Updated: Jul 2, 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

533
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
Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images
08:20

Author Spotlight: AI-Driven Trypanosome Species Detection from Microscopic Images

Published on: October 27, 2023

1.4K

Area of Science:

  • Computer Science
  • Artificial Intelligence
  • Machine Learning

Background:

  • Deep learning models are often used directly for classification by retraining their final layers.
  • Alternative approaches involve using pre-trained models as feature extractors within other classification frameworks.

Purpose of the Study:

  • To investigate if using pre-trained deep learning models as feature extractors in Bag-of-Deep-Visual-Words (BoDVW) consistently improves classification accuracy compared to direct classification.
  • To analyze the impact of various factors, including feature extraction methods, feature encoding techniques, and model fine-tuning strategies.

Main Methods:

  • Employed five feature encoding methods: hard-voting, soft-voting, locally constrained linear coding, super vector coding, and Fisher Vector (FV).
  • Utilized two feature extraction methods: Ext-DFs(CP) (convolutional/non-global pooling layers) and Ext-DFs(FC) (fully-connected/global pooling layers).
  • Experimented with three pre-trained models (VGGNet-16, ResNext-50, Swin-B) on six diverse datasets.

Main Results:

  • Using pre-trained models as feature extractors with Fisher Vector encoding improved accuracy in 35 out of 36 experiments compared to retraining only the classification layer.
  • Accuracy gains ranged from 0.13% to 8.43% (average 3.11%) with Ext-DFs(CP) and 1.06% to 14.63% (average 5.66%) with Ext-DFs(FC).
  • Fine-tuning all layers and using FV with Ext-DFs(FC) improved accuracy in 14 out of 18 experiments, with gains from 0.21% to 5.65% (average 1.58%).

Conclusions:

  • Utilizing pre-trained deep learning models as feature extractors within the BoDVW framework is a promising technique for enhancing classification accuracy.
  • The effectiveness depends on factors like feature encoding (FV being highly effective) and extraction methods, but it does not universally guarantee improved accuracy.
  • This approach offers significant potential for accuracy improvements in various computer vision tasks.