This PDF file contains the front matter associated with SPIE Proceedings Volume 13060, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Muscular myopathies such as woody or Wooden Breast (WB), which impair the eating quality and marketability of poultry products, are threatening the profitability of poultry industries worldwide, with an estimated annual loss exceeding $500 million for the United States (U.S.) poultry industry. WB-affected fillets are characterized by abnormal tissue hardness and muscle rigidity with varying degrees of severity. The assessment of WB conditions at processing facilities currently relies on tactile palpation combined with a visual examination by trained personnel. This approach is subjective, labor-intensive, costly, and may induce contamination due to physical contact. Optical imaging technology offers a promising alternative for objective and non-invasive quality assessment of broiler meat. This study presents a proof-of-concept evaluation of a new scattering imaging technique that captures light-scattering characteristics of meat tissues for the detection of WB conditions in broiler breast fillets. Broadband scattering images, generated under the illumination of a highly focused broadband beam, were acquired from broiler meat samples. Two types of image features, i.e., 1) deep-learning-based and 2) hand-crafted scattering features, were extracted for building classification models using regularized linear discriminant analysis to differentiate meat samples into two categories, i.e., “Normal (no WB)” and “Defective”, according to WB conditions. Deep-learning-based features yielded an overall classification accuracy of 80.9%, while an improved accuracy of 88.7% was obtained by hand-crafted scattering features, representing a significant improvement of 7.8% (P ⪅ 0.01). Furthermore, feature selection based on Minimum Redundancy Maximum Relevance (MRMR) was conducted to select a subset of scattering image features for discriminant modeling, leading to a further accuracy improvement to 90.5% with top-ranked 65 features. This study has demonstrated the promise of the light scattering imaging technique for WB detection in broiler breast meats.

Citrus Black Spot (CBS) disease, caused by the pathogenic fungus Phyllosticta citricarpa, presents a significant threat to citrus-growing regions, including Florida. Detecting CBS early is crucial, especially when trees don't yet show symptoms. This early stage provides an opportunity for orchard managers to take preventive measures and curb the disease's spread. In our study, we explore the CSI-D+ system, which combines cutting-edge fluorescence imaging technology with the YOLOv8 deep learning framework. We focus on identifying two CBS fungus variants, GC12 and GM33, commonly found on infected citrus leaves. Sample leaves were inoculated with varying concentration levels of the two variants and imaged by the CSI-D+ device. Impressively, the CSI-D+ system demonstrates exceptional discrimination abilities for discerning variant concentration levels. It achieves a notable mean accuracy of 96.97% for detecting the GC12 fungus, with an F1- score of 96.35% and a mean average precision (mAP) of 97.82%. Similarly, for the GM33 variant, the system maintains an average accuracy of 96.17%, an F1-score of 88.76%, and a mAP of 91.64%. The system offers promise as rapid, non-invasive tool for early CBS fungus spore detection on citrus leaves. By providing timely insights, it could empower effective intervention strategies, bolstering orchard resilience against this harmful fungus.

As the population of the earth grows, the demand for food grows proportionally. Early and cost-effective detection of plant diseases can result in less food loss throughout the world. The current methods for image-based plant disease detection tend to fail in field conditions. Our method uses region proposal networks to localize diseased leaves for detection. We discard no prior anchor boxes, which increases the average recall of the network, resulting in better localization.

Hyperspectral imaging records data over a broad range of electromagnetic spectrum wavelengths and presents a viable option for fruit maturity detection when incorporated with deep neural networks. This paper focuses on improving the accuracy of the Kiwi and Avocado fruit hyperspectral dataset by introducing a modified version of depthwise separable convolution and comparing the results with state-of-the-art models to prove our model’s reliability. The research aims to use the proposed model to predict the fruits’ ripeness, firmness, and sugar content levels.

The transmission of Escherichia coli (E. coli) bacteria to humans through infected fruits and vegetables, such as citrus, can lead to severe health issues, including bloody diarrhea and kidney disease (Hemolytic Uremic Syndrome). Therefore, the implementation of a suitable sensor and detection approach for inspecting the presence of E. coli colonies on fruits and vegetables would greatly enhance food safety measures. This article presents an evaluation of SafetySpect's Contamination, Sanitization Inspection, and Disinfection (CSI-D+) system, comprising an UV camera, an RGB camera, and illumination at fluorescence excitation wavelengths: ultraviolet C (UVC) at 275 nm. To conduct the study, eight different concentrations ranging from 10⁰ (control) to 10⁸ (maximum) cell counts of bacterial populations were inoculated on extracted citrus peel specimens. Specimen data could represent either irrigation or sprayer-based contamination events or direct contact with wildlife. Our study delves into early detection using the portable CSI-D+ system, capturing 240x240 pixel UV-C fluorescence images of E. Coli-inoculated grapefruit peel plugs. We developed a pipeline to prepare these images for the YOLOv8 deep learning framework, facilitating E. coli classification across varying concentrations and backgrounds. To enhance explainability, we employed Eigen Class Activation Map (Eigen-CAM) with YOLOv8, utilizing 'pytorch-eigen-cam' (https://github.com/rigvedrs/YOLO-V8-CAM) to elucidate the model's decision-making in detecting and classifying different E. coli concentrations. Our study demonstrated that the CSI-D+ system could classify fluorescence images at eight different concentration levels with an overall accuracy of more than 83% in which the control class reached a perfect classification accuracy while the images with E. coli concentration of 10⁶ CFU/drop had the lowest accuracy of 71%. Similarly, the images with maximum concentration i.e., 10⁸ CFU/drop were classified at an accuracy of 94%. These findings demonstrate the application of the CSI-D+ system as a rapid, non-invasive tool for E. coli detection on citrus peel surfaces that may be on the tree thus alerting the potential for similar contamination on fruit still on the tree. By providing timely insights, these results could enable effective intervention strategies to eliminate dangerous E. Coli from the food chain.

Nowadays the Clostridium detection in milk for the dairy industry still is a challenging problem since traditional methods are time-consuming and lack specificity towards these bacteria. The use of microbiological techniques is possible but is expensive in terms of response time and requires qualified personnel. Pasteurization is ineffective against Clostridium spores which can survive the process and later revert to their vegetative form during cheese aging. The Clostridium metabolism is characterized by the production of carbon dioxide and hydrogen, which can lead to the formation of cracks and slits in the cheese altering its taste and structure. The analysis of gas production is indicative of the presence of Clostridia; therefore, it can be exploited to detect their presence. This study presents a Raman spectroscopy-based instrument for a rapid and cost-effective identification of Clostridium in milk. The methodology relies on the widely adopted Most Probable Number (MPN) method, as established by Brändle et al. (2016). However, our innovation lies in adoption of a Raman-based instrument to speed up the vial positivity detection. The instrument also enables the discrimination Clostridia infection from non-hydrogen-producing bacteria.

In the realm of food safety, the standard practice involves collecting food product samples and sending them to a central laboratory for microbiological testing. However, this process introduces delays in obtaining the microbiological testing results and subsequently affects the timely delivery of food products to consumers. To further reduce the time-to-detection issue, we propose the development of a self-contained, battery-operated, high-sensitivity optical sensor that can be affixed to the cap of the typical food sample collection container. This device, called MPACT, offers real-time and in-transit monitoring of the contamination status of the food sample, specifically targeting E. coli O157:H7, through a bioluminescence assay. The assay exclusively targets the target pathogen and, when detected, produces minimal luminescence. As the sample is transported in the container, the number of bacterial cells multiplies, and once the luminescent signal reaches a predefined threshold, the sensor reports the results via Bluetooth. This study focuses on the design of the bottle cap sensor and examines its sensitivity by subjecting it to bioluminescence samples.

Celiac disease is a serious gluten-sensitive autoimmune disease of the small intestine affecting genetically susceptible individuals worldwide. A strict, lifelong gluten-free diet is the only treatment. Currently, the most commonly used methods for gluten test are based upon enzyme-linked immunosorbent assay, which is sample-destructive, and requires cumbersome processing procedures, and therefore are not suitable for high-throughput real-time screening detection of gluten in foods. In this study, a Fourier-Transform Infrared (FT-IR) spectroscopy-based approach was proposed for authentication of gluten-free flour. Three chemical standards including gliadin, gluten, and starch from wheat and 62 different types of flour products were scanned by FT-IR spectroscopy over the wavenumber range of 4000 and 400 cm^-1. Notable absorbance differences were observed between the chemical standards of gliadin and gluten and starch from wheat over the wavenumber range of 1800-450 cm^-1. The mean absorbance profiles of gluten-free and non-gluten free categories of flour demonstrated varying spectral characteristics between 1800 and 1500 cm^-1. The Principal Component Analysis (PCA)- Linear Discriminant Analysis (LDA) models built upon the original absorbance of flour between 1800 and 1500 cm^-1 achieved overall prediction accuracies of at least 95.7%. The potential of FT-IR technique in identifying and authenticating gluten-free flour was demonstrated.

With the development and expansion of the internet of things, many scientific and engineering instruments are leaving the benchtop restriction and moving on to provide on-site detection. On-site detection requires a complete miniaturization of a benchtop system while maintaining a similar performance with respect to the analyte detection sensitivity. In addition, due to the mobile nature, utilizing a battery source is required. Here we present a portable loop-medicated isothermal amplification detection system for on-site detection and amplification of target analyte via fluorescence detection. The digital twin design incorporates three major components: an isothermal heating chamber, light-tight enclosure for sample insert, and fluorescence imaging system via micro-controllers. The isothermal heating chamber was designed with Peltier heater to provide small form factor accurate temperature control. For light-tight enclosure is a 3D printed device that allows DNA samples to be inserted and fluorescent images to be taken within the chamber. Lastly, fluorescent imaging system operates with a stand-alone camera connected to an Arduino micro-controller. Excitation is provided by blue colored LED and emission is detected via long-pass filter that matches the emission spectrum.

The rapid and nondestructive identification of pork spoilage holds significant importance due to the inherent richness of nutrients and the conducive environment for bacterial proliferation within pork. This study focused on the non-destructive assessment of the total viable count in pork utilizing visible/near-infrared spectroscopy. By employing this technique, pork samples were subjected to analysis across the visible/near-infrared spectra range (400-1000 nm), with the total viable count determined through the plate counting method. The principal component analysis technique was used to consider whether there was variability in the spectra of pork with different total viable counts. Three different preprocessing methods in visible near-infrared spectroscopy for the prediction of total viable count in pork were compared, the preprocessing methods used were standard normal variate, multiplicative scatter correction and Savitzky-Golay smoothing. The results of the study show that, divisibility of pork with different total viable count in the low-dimensional space of the first and second principal components of principal component analysis. Among these preprocessing techniques, the study highlighted the superiority of the partial least squares regression model combined with standard normal variate preprocessing. This optimized model exhibited remarkable efficiency in predicting total viable count in pork. The best total viable count prediction model showed the R_P, RMSEP, and RPD of 0.864, 0.826%, and 1.887, respectively. This study highlights the importance of rapid and non-destructive techniques for pork spoilage detection, contributing to improved food safety and quality assurance practices within the pork industry.

Excessive consumption of β-adrenergic agonists from livestock, poultry, or viscera can present serious health risks, potentially endangering lives. Surface-enhanced Raman scattering (SERS) provides a precise method for measuring levels of β-adrenergic agonists. This study involved collecting spectra of ractopamine aqueous solutions by synthesizing Au@Ag NPS alloy substrates. A linear relationship between the concentration of ractopamine (ranged from 1 to 10 mg/L) and SERS intensity. Automatic Whittaker Filter (AWF) algorithm was used to preprocess the Raman spectra to remove the fluorescence background. A linear regression model was established between the SERS intensity of different Raman characteristic peaks of ractopamine and the content of ractopamine solution. The established model had linear relationship with a correlation coefficient R2 of 0.98 and RMSE of 0.332 mg/L. This method provides a new idea for the determination of ractopamine. This study is helpful to develop a simple, low-cost and easy-to-store SERS method for the detection of ractopamine based on Au@Ag NPS.

Establishing a universal and efficient method for determining ractopamine residues in pork is of paramount importance for ensuring food safety. However, the main challenge lies in achieving accurate quantitative detection of complex samples using Surface-Enhanced Raman Scattering (SERS), as it requires overcoming interference from substrate-sample mixing and variations in hotspot distribution. This study introduces an innovative approach to address this challenge by proposing a breakthrough interference factor removal network based on deep learning, termed SERSNet. By enhancing the depth of SERS spectroscopy, SERSNet establishes a correlation between the spectra of pork samples with varying concentrations of ractopamine. A multilayer convolution module is developed to effectively extract the spectral features of ractopamine. The Mean Absolute Error (MAE), root mean square error (RMSE), and Mean Absolute Percentage Error (MAPE) of the proposed model in this paper are 0.90, 0.48, and 80.48, respectively. The performance of the SERSNet model surpasses that of the Multiple Linear Regression (MLR) model. The SERSNet algorithm proposed in this paper demonstrates competitiveness and yields superior results.

The SERS technique has great potential for the rapid detection of foodborne microorganisms. However, the rapid detection of Salmonella in pork is difficult because of the complex chemical composition contained in the food matrix. In order to resolve more effective data information from the Raman spectra of complex samples, the spectral data were firstly subjected to data preprocessing such as smoothing, de-baselining, and feature extraction. And the optimal number of windows for SG smoothing was explored. Then, the study performed PLSR analysis on three groups of extracted Raman feature: peak intensities, widths, and area of regions. Besides, the three groups of features were screened using cars. The modeling results showed that compared to the single group of features, the prediction of the model built by the Raman intensity at the 670 cm^-1 , 737 cm^-1 , 804 cm^-1 , 896 cm^-1 , 1096 cm^-1 , 1231 cm^-1 , 1377 cm^-1 , and 1603 cm^-1 Raman shifts, the widths of the characteristic peaks at the 1096 cm^-1 and 1341 cm^-1 shifts, and the areas of peak regions at the 737 cm^-1 , 804 cm^-1 , 1096 cm^-1 , 1341 cm^-1 , and 1377 cm^-1 was improved to a large extent. A predictive correlation of 0.9058 and a predictive root mean square error of 0.9706 were achieved. The method mined more data related to the parameters to be measured from the Raman mapping features and provided a methodological reference for the SERS mapping resolution of complex samples.

The Fish Supply Chain (FSC) industry faces a significant challenge in efficiently and affordably preserving fish quality and detecting adulteration throughout the chain. Quality, Adulteration and Traceability (QAT) is a multi-mode spectroscopy and AI-based handheld device that is developed by our team to identify fish species and assess fish freshness that can be integrated into the FSC ecosystem. We conducted a survey interviewing professionals across the FSC, including harvesters, processors, distributors, and retailers and queried them about how they evaluate fish freshness and the major issues faced in freshness inspection and fraud detection. We learned that traditional sensory evaluation and electronic noses are the most common methods used for fish quality and freshness assessment. QAT technology will play a role as a substitute for current methods and will offer rapid results for fish species identification, quality assessment, and nutritional content analysis. Blockchain (BC), as a Distributed Ledger Technology (DLT), can be integrated with FSC to securely monitor and record fish quality and freshness values each step of the FSC. This helps in maintaining product integrity and provides stakeholders with access to the entire journey of the fish product. We extend our experiments to study the degradation of fish freshness throughout the FSC to trigger the system once the rate of decay exceeds a certain limit. These results should be used so BC integration with smart contracts be able to compare its freshness grade to the history of recorded values. If the degradation in freshness exceeds the expected range, then the smart contract should raise an alarm to alert the system. In this way, BC-based FSC incorporating QAT technology is able to detect any degradation and flag products that may have compromised freshness or quality. This integration of technologies not only promises to revolutionize the FSC but also addresses issues like fraud and illegal fishing activities, ultimately delivering higher-quality and more transparent fish products to consumers.

We utilized hyperspectral imaging technology, which is commonly used for nondestructive quality assessment in agriculture, to predict SSC (Brix, %) and also the firmness (N) of apples. In this research, various regression models were applied based on machine learning and deep learning with hyperspectral (400~1000 nm) spectrum data to predict SSC and firmness of apple fruits. To evaluate the prediction accuracy of each model, coefficient of determination (r square) and Root Mean Square Error (RMSE) was used. For this purpose, spectral data of apple fruits was acquired and prediction models using various regression models such as PLSR were developed. Also, various preprocessing methods were applied, including extracting meaningful pixels, MSC (Multiplicative Scatter Correction), SNV (Standard Normal Variate), to enhance the accuracy of regression models. Through these process, SSC and firmness prediction performance of each model was analyzed and compared with various combination of preprocessing methods.

Salmonella ser. Typhimurium is notorious for causing serious foodborne illnesses and presenting considerable public health risks. The study introduces an innovative system based on a quartz crystal microbalance, designed to detect the target pathogen by integrating the system around a smartphone. The system operates through a dual-mode approach, relying on two distinct mechanisms: measuring frequency changes due to variations in bacterial mass and quantifying fluorescence intensities resulting from bacteria captured by FITC-labeled antibodies. Incorporating FITC-labeled antibodies not only enhances the resonance frequency shift but also offers visual confirmation through the fluorescence signal. The integration of the quartz crystal microbalance system with a smartphone enables real-time monitoring. This system displays both frequency and temperature data, while also capturing fluorescence intensities to estimate the concentration of the target analyte. The smartphone-based system successfully detected Salmonella Typhimurium within a concentration range of 10⁵ CFU/mL after the application of FITC-labeled antibodies. This portable QCM system represents a promising advancement in pathogen detection, holding significant potential to improve food safety protocols and strengthen public health safeguards.

The Terahertz (THz) spectrum, situated between microwave and infrared frequencies, is utilized for applications such as imaging and chemical analysis. This project aims to combine THz sensing with the Kalman Filter estimator, a recursive algorithm typically utilized for estimating data with high levels of noise. Upon a thorough effort, we investigate the prospects and challenges in THz sensing using the Kalman filter for assistance. Upon examining this area, we found an increasing pattern in combining THz wavelength technologies that boosts the precision and effectiveness of Kalman Filter applications in THz sensing. The topic being talked about is the significance of this technology in improving the dependability of detecting and measuring data in difficult conditions. Moreover, this document outlines essential advancements and trends in the field of THz sensing, as well as the progress in various disciplines that contribute to the development of THz sensing. This paper presents details about the current planning and disadvantages of utilizing Kalman Filters in THz sensing. It has served as a broad guide for both scientists and professionals, laying the groundwork for future advancements in this field. Incorporating these directivity methods not only enhances detection capabilities but also produces novel contributions to the field of THz sensing that also facilitate knowledge sharing.

This paper investigates the integration of Kalman Filter technologies into modern agricultural practices, with a focus on advancing crop monitoring techniques. Through a comprehensive bibliometric and text mining analysis, we explore the evolution of research in this area from 1983 to 2023, highlighting the steady growth in scholarly publications and the interdisciplinary nature of the field. Our findings reveal a significant interest in leveraging Kalman Filter algorithms to enhance the precision and efficiency of agricultural operations, addressing the challenges of sustainability and food security. The study underscores the importance of data assimilation, real-time monitoring, and predictive analysis in agriculture, facilitated by the adoption of Kalman Filter and its variants. We also examine the integration of emerging technologies, such as Unmanned Aerial Vehicles (UAVs) and remote sensing, with Kalman Filter techniques to develop sophisticated agricultural monitoring systems. The paper concludes with insights into future research directions, emphasizing the need for overcoming barriers to technology adoption and fostering interdisciplinary collaborations. This work contributes to the understanding of how computational intelligence can transform agricultural practices, offering solutions for more sustainable and efficient farming.