Skip to main content
ARS Home » Southeast Area » Athens, Georgia » U.S. National Poultry Research Center » Quality and Safety Assessment Research Unit » Research » Publications at this Location » Publication #411435

Research Project: Smart Optical Sensing of Food Hazards and Elimination of Non-Nitrofurazone Semicarbazide in Poultry

Location: Quality and Safety Assessment Research Unit

Title: 3D-GhostNet: A novel spatial-spectral algorithm to improve foodborne bacteria classification coupled with hyperspectral microscopic imaging technology

item KANG, RUI - Jiangsu Academy Agricultural Sciences
item SUN, SHANGPENG - McGill University - Canada
item OUYANG, QIN - Jiangsu University
item HUANG, JIAXING - Jiangsu Academy Agricultural Sciences
item Park, Bosoon

Submitted to: Sensors and Actuators B: Chemical
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/24/2024
Publication Date: 4/1/2024
Citation: Kang, R., Sun, S., Ouyang, Q., Huang, J., Park, B. 2024. 3D-GhostNet: A novel spatial-spectral algorithm to improve foodborne bacteria classification coupled with hyperspectral microscopic imaging technology. Sensors and Actuators B: Chemical.

Interpretive Summary: Foodborne pathogens are crucial precipitators of food poisoning and foodborne illnesses, posing a severe threat to consumer health and safety. Approximately 600 million people fall ill after eating contaminated food each year, resulting in 420,000 deaths and the loss of 33 million healthy life years worldwide. However, current methods for pathogen detection are time-consuming, often missing the critical window for controlling and preventing outbreaks of foodborne epidemics. Therefore, the development of early and rapid detection techniques for foodborne pathogens remains one of the global challenges in food safety. In this research, we developed innovative technologies using imaging technology integrated with artificial intelligence (AI) data analytics needed by the food industry and regulatory agencies for rapidly and accurately detecting food safety problems. Our AI-based data processing and analysis methods can classify complex high-dimensional hyperspectral image data acquired from various foodborne bacteria with 100% accuracy that surpassed previous results from the conventional foodborne detection methods.

Technical Abstract: A novel 3D-GhostNet was developed for hyperspectral microscopic imaging (HMI) data analytics to improve identification of foodborne pathogens. The HMI, leveraging microscopic technology, surpasses traditional spectral imaging in terms of sensitivity and resolution by utilizing visible/near-infrared spectroscopy integrated with high-resolution single-cell images. The newly constructed 3D-Ghost network directly processes and recognizes single-cell hypercube, enables HMI technology to become near-real-time detection by combining with automated data processing pipeline. 3D-GhostNet incorporates Ghost modules as its backbone and adds convolutional block attention module (CBAM) for adaptively extracting high-dimensional depth spatial-spectral information. In the task of identifying four different foodborne bacteria cells, our 3D-GhostNet achieved 100% accuracy, surpassing spectral-only-based (SOB) classifiers such as linear discriminant analysis (LDA, 89.7%), support vector machine (SVM, 93.9%), one-dimensional convolutional neural network (1D-CNN, 98%), and long short-term memory (LSTM, 98.5%). Furthermore, our independent blind test results indicate that 3D-GhostNet paradigm remains robust for identifying complex mixtures of pathogens. As a label-free detection tool, 3D-GhostNet-assisted HMI technology is fully competent in classification of different foodborne pathogens, suggesting broad applications in food safety and quality.