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

Classification of Different Fermentation Stages of Black Tea Using a Lightweight CNN Optimized by Knowledge

Xuteng Liu1,2, Mengqi Guo1, Zhengtong He1

  • 1Tea Research Institute of Shandong Academy of Agricultural Sciences, Jinan 250100, China.

Foods (Basel, Switzerland)
|May 27, 2026
PubMed
Summary

Related Concept Videos

Classification of Systems-I01:26

Classification of Systems-I

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:
Classification of Systems-II01:31

Classification of Systems-II

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,

You might also read

Related Articles

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

Sort by
Same author

FSCA-EUNet: Lightweight Classification of Stacked Jasmine Bloom-Stages via Frequency-Spatial Cross-Attention for Industrial Scenting Automation.

Foods (Basel, Switzerland)·2025
Same author

Assessment of Tenderness and Anthocyanin Content in Zijuan Tea Fresh Leaves Using Near-Infrared Spectroscopy Fused with Visual Features.

Foods (Basel, Switzerland)·2025
Same author

Integrated Microbiology and Metabolomics Analysis Reveal Responses of Soil Bacterial Communities and Metabolic Functions to N-Zn Co-Fertilization in the Rhizosphere of Tea Plants (<i>Camellia sinensis</i> L.).

Plants (Basel, Switzerland)·2025
Same author

ITD-YOLO: An Improved YOLO Model for Impurities in Premium Green Tea Detection.

Foods (Basel, Switzerland)·2025
Same author

Predicting the Degree of Fresh Tea Leaves Withering Using Image Classification Confidence.

Foods (Basel, Switzerland)·2025
Same author

Effect of withering/spreading on the physical and chemical properties of tea: A review.

Comprehensive reviews in food science and food safety·2024

This study introduces AT-ShuffleNet, a lightweight AI model for accurately identifying red tea fermentation stages. This automated approach improves tea quality and reduces production losses.

Area of Science:

  • Agricultural Engineering
  • Computer Vision
  • Food Science

Background:

  • Fermentation is crucial for red tea flavor, but manual stage determination is imprecise and inefficient.
  • Inaccurate fermentation control leads to flavor loss and reduced product value in red tea production.

Purpose of the Study:

  • To develop an accurate and efficient automated system for identifying red tea fermentation stages.
  • To create a lightweight classification model for real-time application in tea processing.

Main Methods:

  • A Convolutional Neural Network (CNN) and knowledge distillation were combined to create the AT-ShuffleNet model.
  • ResNet and ShuffleNet v2-0.5 were used as teacher and student models, respectively, for knowledge distillation.
  • Focal and Poly Losses, along with various distillation strategies (STD, MGD, SPKD, ATD, KD), were employed to optimize the model.
Keywords:
CNN modeledge deploymentfermentation classificationknowledge distillationlightweight

Related Experiment Videos

Main Results:

  • The AT-ShuffleNet model achieved high performance metrics: Precision (89.11%), Recall (90.16%), Kappa (89.29%), Accuracy (91.2%), and F1-score (89.53%).
  • The lightweight model demonstrated effectiveness in accurate classification and addresses limitations of manual assessment.
  • Industrial validation included deployment on edge devices and integration into a WeChat mini-program.

Conclusions:

  • The developed AT-ShuffleNet model offers a viable solution for accurate and efficient red tea fermentation stage identification.
  • This automated system can significantly reduce manual intervention, improve product quality, and minimize economic losses in the tea industry.
  • The model's lightweight nature and successful edge deployment facilitate practical application in diverse, unstructured processing environments.