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

You might also read

Related Articles

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

Sort by
Same author

Parameter- and Compute-Efficient Spatial-Spectral Transformer Framework for Pixel-Level Classification of Foreign Plastic Objects on Broiler Meat Using NIR-Hyperspectral Imaging.

Sensors (Basel, Switzerland)·2026
Same author

Combined Relaxation Spectra for the Prediction of Meat Quality: A Case Study on Broiler Breast Fillets with the Wooden Breast Condition.

Foods (Basel, Switzerland)·2024
Same author

Development of Multimodal Fusion Technology for Tomato Maturity Assessment.

Sensors (Basel, Switzerland)·2024
Same author

Semisupervised Deep Learning for the Detection of Foreign Materials on Poultry Meat with Near-Infrared Hyperspectral Imaging.

Sensors (Basel, Switzerland)·2023
Same author

Detection of Aflatoxin B<sub>1</sub> in Single Peanut Kernels by Combining Hyperspectral and Microscopic Imaging Technologies.

Sensors (Basel, Switzerland)·2022
Same author

Development of Imaging System for Online Detection of Chicken Meat with Wooden Breast Condition.

Sensors (Basel, Switzerland)·2022

Related Experiment Video

Updated: May 28, 2025

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

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

8.9K

Deep Learning Model Compression and Hardware Acceleration for High-Performance Foreign Material Detection on Poultry

Zirak Khan1, Seung-Chul Yoon2, Suchendra M Bhandarkar1

  • 1School of Computing, University of Georgia, Athens, GA 30602, USA.

Sensors (Basel, Switzerland)
|February 13, 2025
PubMed
Summary
This summary is machine-generated.

Optimizing deep learning inference for hyperspectral imaging (HSI) with hardware acceleration and quantization significantly speeds up foreign material detection in poultry processing. This enables real-time quality control at high line speeds, ensuring safer food products.

Keywords:
deep learningforeign material detectionhardware accelerationhyperspectral imagingindustrial applicationsinference optimizationpost-training quantizationpoultry processingreal-time processingsensor technology

More Related Videos

Lensless Fluorescent Microscopy on a Chip
11:23

Lensless Fluorescent Microscopy on a Chip

Published on: August 17, 2011

17.6K
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

8.5K

Related Experiment Videos

Last Updated: May 28, 2025

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

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

8.9K
Lensless Fluorescent Microscopy on a Chip
11:23

Lensless Fluorescent Microscopy on a Chip

Published on: August 17, 2011

17.6K
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

8.5K

Area of Science:

  • Food science and technology
  • Computer vision
  • Machine learning

Background:

  • Food safety relies on detecting foreign materials in poultry.
  • Hyperspectral imaging (HSI) provides rich data for contaminant detection.
  • Deep learning (DL) models offer high accuracy but face real-time processing challenges.

Purpose of the Study:

  • To optimize deep learning inference for hyperspectral imaging-based foreign material detection in poultry processing.
  • To address challenges of high dimensionality, computational complexity, and real-time implementation.

Main Methods:

  • Applied post-training quantization (FP16) to reduce model size and computational load.
  • Utilized hardware acceleration with NVIDIA TensorRT for enhanced inference speed.
  • Simulated hyperspectral line-scan cameras for industrial condition evaluation.

Main Results:

  • Achieved inference times compatible with poultry processing line speeds (140-250 birds/min).
  • Reduced inference time by up to five times compared to traditional GPU inference.
  • Decreased model size by 50% while maintaining high detection accuracy.

Conclusions:

  • Integration of post-training quantization and hardware acceleration is effective for real-time DL inference on HSI data.
  • Optimized models show potential for practical deployment in industrial poultry processing.
  • This approach overcomes computational bottlenecks for improved food quality and safety.