Incorporating nanomaterials into food products provides key benefits, including extended shelf life, improved safety, and enhanced quality and texture. These innovations could help tackle major challenges in modern food systems, such as reducing waste and enhancing food quality and safety. However, potential toxicity remains a concern, compounded by the lack of physiologically relevant models for assessing ingested nanomaterials. Traditional in vitro and in vivo approaches often fail to mimic gastrointestinal complexity, resulting in inconsistent and non predictive nanotoxicity data that hinder accurate risk assessment of nano enabled foods. To address this gap, this review evaluates the potential of microphysiological systems (MPS), particularly gut-targeted MPS, for modeling gastrointestinal nanoparticle exposure. It examines how MPS technologies replicate key physiological processes relevant to food specific risk assessment, including intestinal barrier function, microbiota immune interactions, and gut organ communication. A comparative analysis of technological advances and their applications in nanotoxicology explores how MPS can be better adapted for nanofood safety evaluation.
Gabriel de Oliveira Quintana, Renata Fragoso Potenza, Sofia Lasmar Lima Oliveira
Brazil is the world’s fifth-largest methane emitter, with methane accounting for 24% of national greenhouse gas emissions and the agricultural sector responsible for 75.6% of these methane emissions, mainly from livestock. Under a business-as-usual scenario, agricultural emissions are projected to rise in line with intrinsic emissions associated with production growth, challenging Brazil’s commitments to the Global Methane Pledge and the Nationally Determined Contribution (NDC). Based on the methodology for calculating emissions from the Brazilian National GHG Inventory and the mitigation potential of different mitigation strategies aimed at the agricultural sector, such as those addressed by the ABC+ Plan. This brief compares a trend scenario against a mitigation scenario based on the adoption of proven low-emission practices and technologies in livestock and crops productions systems. The implementation of methane reduction strategies for livestock and agricultural production can reduce sectoral methane emissions by 28% by 2035 compared to 2020 levels. Policymakers must prioritize scaling of these technologies, replacing the more methane-emitting analogues and establish robust Monitoring, Reporting, and Verification (MRV) systems to ensure impact.
Nutrition. Foods and food supply, Food processing and manufacture
With rising levels of obesity and associated non-communicable diseases in Africa, there is an increasing concern with regards to ultra-processed foods. Whilst regulation of these foods is important, it can lead to the demonization of all forms of processed foods. However, healthy processed and minimally processed foods play a crucial role in reducing food loss and waste, improving food safety, and providing scarce nutrients otherwise inaccessible or unaffordable. Further potential lies in improving gender equality, and boosting economic opportunities through a growing agro-processing sector. A nuanced approach is therefore needed, leveraging the opportunities of (minimally) processed foods whilst discouraging sales and consumption of ultra-processed foods.
Food image recognition is a challenging task in computer vision due to the high variability and complexity of food images. In this study, we investigate the potential of Noisy Vision Transformers (NoisyViT) for improving food classification performance. By introducing noise into the learning process, NoisyViT reduces task complexity and adjusts the entropy of the system, leading to enhanced model accuracy. We fine-tune NoisyViT on three benchmark datasets: Food2K (2,000 categories, ~1M images), Food-101 (101 categories, ~100K images), and CNFOOD-241 (241 categories, ~190K images). The performance of NoisyViT is evaluated against state-of-the-art food recognition models. Our results demonstrate that NoisyViT achieves Top-1 accuracies of 95%, 99.5%, and 96.6% on Food2K, Food-101, and CNFOOD-241, respectively, significantly outperforming existing approaches. This study underscores the potential of NoisyViT for dietary assessment, nutritional monitoring, and healthcare applications, paving the way for future advancements in vision-based food computing. Code for reproducing NoisyViT for food recognition is available at NoisyViT_Food.
Amir Hosseinian, Ashkan Dehghani Zahedani, Umer Mansoor
et al.
Progress in AI for automated nutritional analysis is critically hampered by the lack of standardized evaluation methodologies and high-quality, real-world benchmark datasets. To address this, we introduce three primary contributions. First, we present the January Food Benchmark (JFB), a publicly available collection of 1,000 food images with human-validated annotations. Second, we detail a comprehensive benchmarking framework, including robust metrics and a novel, application-oriented overall score designed to assess model performance holistically. Third, we provide baseline results from both general-purpose Vision-Language Models (VLMs) and our own specialized model, january/food-vision-v1. Our evaluation demonstrates that the specialized model achieves an Overall Score of 86.2, a 12.1-point improvement over the best-performing general-purpose configuration. This work offers the research community a valuable new evaluation dataset and a rigorous framework to guide and benchmark future developments in automated nutritional analysis.
Chinedu Emmanuel Mbonu, Kenechukwu Anigbogu, Doris Asogwa
et al.
Food recognition systems has advanced significantly for Western cuisines, yet its application to African foods remains underexplored. This study addresses this gap by evaluating both deep learning and traditional machine learning methods for African food classification. We compared the performance of a fine-tuned ResNet50 model with a Support Vector Machine (SVM) classifier. The dataset comprises 1,658 images across six selected food categories that are known in Africa. To assess model effectiveness, we utilize five key evaluation metrics: Confusion matrix, F1-score, accuracy, recall and precision. Our findings offer valuable insights into the strengths and limitations of both approaches, contributing to the advancement of food recognition for African cuisines.
Abstract Fermented plant foods, deeply rooted in cultural traditions, are gaining increasing attention for their potential to modulate the gut microbiome and improve host health. This review summarizes current data on the microbial composition, functional metabolites, and health effects of fermented plant foods commonly consumed in Asia, with a focus on fermented soybean foods (e.g., cheonggukjang, natto, and tempeh), and fermented vegetable foods (e.g., kimchi). Several bioactive compounds derived from fermentation modulate gut microbial composition and diversity, gut barrier integrity, and immune and inflammatory responses to help prevent and manage metabolic disorders, inflammatory bowel disease, and other gut-related disorders. Preclinical and animal studies have elucidated the mechanisms underlying these health effects. We highlight the importance of developing personalized dietary interventions, standardizing the production of fermented plant foods, and evaluating health effects using multi-omics approaches. These foods hold promise as microbiome-targeted interventions for maintaining and improving host health.
Nutrition science and food science & technology are crucial for creating a healthier world through accessible nutrition and sustainable health practices. Examples of successful impact can be found in food fortification, foods with effective levels of bioactives, (re)formulation of foods to combat obesity and diet-related diseases, (re)formulation of foods to enable nutrition and health claims, and the activities by the European Technology Platform (ETP) for the food sector 'Food for Life'.In preparing, maintaining, and promoting the Strategic Research and Innovation Agenda (SRIA), the ETP aims to identify scientific and technological actions towards the transformations that are needed to achieve more optimal outcomes of the food system. The four major food system outcomes are the environment, society, citizen health & wellbeing, and economy & competitiveness. The SRIA provides essential guidance to the European Commission, Member States and Regions, the food industry, and the wider research community interested in food, in the form of Research and Innovation needs to make a real difference to the Food and Drink sector and society. The mutualism between nutritionists and food scientists and technologists is essential for achieving the transformations towards the outcomes that are needed for a more sustainable food system.
Nutrition. Foods and food supply, Food processing and manufacture
Evripidis P. Kechagias, Sotiris P. Gayialis, Georgios A. Papadopoulos
et al.
In today’s era, humanity has been overwhelmed by technological revolutions that have changed and will continue to change how business operations are performed, directly or indirectly. At the same time, the processes within the supply chain are quite complex, and as technology and processes evolve, they become more and more challenging. Traceability has become a critical issue in the food industry to ensure safety, quality, and compliance with regulations. The adoption of blockchain technology in the food supply chain has gained significant attention as a potential solution to improve traceability. This paper presents the development of a distributed application for table olives’ traceability on the Ethereum network. The paper also presents a methodological framework, which can help anyone aiming to implement an Ethereum decentralized application and demonstrates the practical use of the developed application by a Greek table olives producer. The application significantly improved the producer’s product traceability by providing a secure, transparent, and efficient solution for tracking and tracing the products in the supply chain. The app reduced the time, increased the accuracy and reliability of data, improved supply chain efficiency, and helped the producer comply with international regulations and standards.
Theofilos D. Mastos, Katerina Gotzamani, Petros Ieromonachou
et al.
This paper presents a model designed to measure and investigate the relationships between critical factors, practices, and performance of sustainable supply chain management (SSCM) in the food industry. A survey of 423 firms in the Greek food industry was conducted to meet these objectives. The data were analyzed using exploratory factor analysis, followed by confirmatory factor analysis and structural equation modeling, to explore the relationships between the model’s constructs. The results indicate that “firm-level critical sustainability factors” and “supply chain critical sustainability factors” significantly enhance “supply chain collaboration” and “supply chain strategic orientation”. Additionally, “supply chain strategic orientation” positively influences “social performance” and “environmental performance”, while “supply chain collaboration” positively affects “economic performance” and “environmental performance”. Furthermore, “social performance” is found to have a significant positive impact on “economic performance”. This study provides empirical evidence that helps managers understand the importance of the relationships among SSCM critical factors, SSCM practices, and SSCM performance, and enables them to allocate resources effectively and design SSCM strategies. Finally, the developed constructs offer a measurement tool useful for both practitioners implementing SSCM in their firms and researchers who wish to apply or test the proposed scales in other projects or use them as benchmarks.
Global food production and trade networks are highly dynamic, especially in response to shortages when countries adjust their supply strategies. In this study, we examine adjustments across 123 agri-food products from 192 countries resulting in 23616 individual scenarios of food shortage, and calibrate a multi-layer network model to understand the propagation of the shocks. We analyze shock mitigation actions, such as increasing imports, boosting production, or substituting food items. Our findings indicate that these lead to spillover effects potentially exacerbating food inequality: an Indian rice shock resulted in a 5.8 % increase in rice losses in countries with a low Human Development Index (HDI) and a 14.2 % decrease in those with a high HDI. Considering multiple interacting shocks leads to super-additive losses of up to 12 % of the total available food volume across the global food production network. This framework allows us to identify combinations of shocks that pose substantial systemic risks and reduce the resilience of the global food supply.
Shayan Rokhva, Babak Teimourpour, Amir Hossein Soltani
In contemporary society, the application of artificial intelligence for automatic food recognition offers substantial potential for nutrition tracking, reducing food waste, and enhancing productivity in food production and consumption scenarios. Modern technologies such as Computer Vision and Deep Learning are highly beneficial, enabling machines to learn automatically, thereby facilitating automatic visual recognition. Despite some research in this field, the challenge of achieving accurate automatic food recognition quickly remains a significant research gap. Some models have been developed and implemented, but maintaining high performance swiftly, with low computational cost and low access to expensive hardware accelerators, still needs further exploration and research. This study employs the pretrained MobileNetV2 model, which is efficient and fast, for food recognition on the public Food11 dataset, comprising 16643 images. It also utilizes various techniques such as dataset understanding, transfer learning, data augmentation, regularization, dynamic learning rate, hyperparameter tuning, and consideration of images in different sizes to enhance performance and robustness. These techniques aid in choosing appropriate metrics, achieving better performance, avoiding overfitting and accuracy fluctuations, speeding up the model, and increasing the generalization of findings, making the study and its results applicable to practical applications. Despite employing a light model with a simpler structure and fewer trainable parameters compared to some deep and dense models in the deep learning area, it achieved commendable accuracy in a short time. This underscores the potential for practical implementation, which is the main intention of this study.
Integrated global food system analysis is hampered by the fragmentation of data among food types, processes, and scales. Studies also often neglect the connection to human metabolism -- the ultimate driver of food demand. Here we use a common energetic framework to harmonize data on 95 individual food commodities across food system processes, including production, processing, animal feed and consumption, and estimate human metabolism from body size, demographic, and activity data. We estimate that the share of unmetabolized food calories globally doubled between 1990 and 2019 (from about 10 to 20% of the total calories available for human consumption) as food supply outpaced energy expenditure. Approximately half (51%) of the global population's metabolic demands could theoretically be met by production in the same local 1-degree grid cell (~ 10,000 km2) when holding diets constant. Our open-source framework can be applied to assess strategies to reduce food system inefficiencies from photosynthesis to metabolism while meeting local energetic demands.
Emmanuel Addai, Delfim F. M. Torres, Zalia Abdul-Hamid
et al.
We propose and analyze a deterministic mathematical model for the transmission of food-borne diseases in a population consisting of humans and flies. We employ the Caputo operator to examine the impact of governmental actions and online food delivery services on the transmission of food-borne diseases. The proposed model investigates important aspects such as positivity, boundedness, disease-free equilibrium, basic reproduction number and sensitivity analysis. The existence and uniqueness of a solution for the initial value problem is established using Banach and Schauder type fixed point theorems. Functional techniques are employed to demonstrate the stability of the proposed model under the Hyers-Ulam condition. For an approximate solution, the iterative fractional order Predictor-Corrector scheme is utilized. The simulation of this scheme is conducted using Matlab as the numeric computing environment, with various fractional order values ranging from 0.75 to 1. Over time, all compartments demonstrate convergence and stability. The numerical simulations highlight the necessity for the government to implement the most effective food safety control interventions. These measures could involve food safety awareness and training campaigns targeting restaurant managers, staff members involved in online food delivery, as well as food delivery personnel.
Traditional dietary assessment methods heavily rely on self-reporting, which is time-consuming and prone to bias. Recent advancements in Artificial Intelligence (AI) have revealed new possibilities for dietary assessment, particularly through analysis of food images. Recognizing foods and estimating food volumes from images are known as the key procedures for automatic dietary assessment. However, both procedures required large amounts of training images labeled with food names and volumes, which are currently unavailable. Alternatively, recent studies have indicated that training images can be artificially generated using Generative Adversarial Networks (GANs). Nonetheless, convenient generation of large amounts of food images with known volumes remain a challenge with the existing techniques. In this work, we present a simple GAN-based neural network architecture for conditional food image generation. The shapes of the food and container in the generated images closely resemble those in the reference input image. Our experiments demonstrate the realism of the generated images and shape-preserving capabilities of the proposed framework.
Leah Costlow, Anna Herforth, Timothy B. Sulser
et al.
Most people around the world still lack access to sufficient quantities of all food groups needed for an active and healthy life. This study traces historical and projected changes in global food systems toward alignment with the new Healthy Diet Basket (HDB) used by UN agencies and the World Bank to monitor the cost and affordability of healthy diets worldwide. Using the HDB as a standard to measure adequacy of national, regional and global supply-demand balances, we find substantial but inconsistent progress toward closer alignment with dietary guidelines, with large global shortfalls in fruits, vegetables, and legumes, nuts, and seeds, and large disparities among regions in use of animal source foods. Projections show that additional investments aimed at reducing chronic hunger would modestly accelerate improvements in adequacy where shortfalls are greatest, revealing the need for complementary investments to increase access to under-consumed food groups especially in low-income countries.
Efficient and comprehensive analysis of lipid profiles in yak ghee samples collected from different elevations is crucial for optimal utilization of these resources. Unfortunately, such research is relatively rare. Yak ghee collected from three locations at different altitudes (S2: 2986 m; S5: 3671 m; S6: 4508 m) were analyzed by quantitative lipidomic. Our analysis identified a total of 176 lipids, and 147 s lipid of them were upregulated and 29 lipids were downregulated. These lipids have the potential to serve as biomarkers for distinguishing yak ghee from different altitudes. Notably, S2 exhibited higher levels of fatty acids (21:1) and branched fatty acid esters of hydroxy fatty acids (14:0/18:0), while S5 showed increased levels of phosphatidylserine (O-20:0/19:1) and glycerophosphoric acid (19:0/22:1). S6 displayed higher levels of triacylglycerol (17:0/20:5/22:3), ceramide alpha-hydroxy fatty acid-sphingosine (d17:3/34:2), and acyl glucosylceramides (16:0–18:0–18:1). Yak ghee exhibited a high content of neutralizing glycerophospholipids and various functional lipids, including sphingolipids and 21 newly discovered functional lipids. Our findings provide insights into quantitative changes in yak ghee lipids during different altitudes, development of yak ghee products, and screening of potential biomarkers.
Nutrition. Foods and food supply, Food processing and manufacture
Manually tracking nutritional intake via food diaries is error-prone and burdensome. Automated computer vision techniques show promise for dietary monitoring but require large and diverse food image datasets. To address this need, we introduce NutritionVerse-Synth (NV-Synth), a large-scale synthetic food image dataset. NV-Synth contains 84,984 photorealistic meal images rendered from 7,082 dynamically plated 3D scenes. Each scene is captured from 12 viewpoints and includes perfect ground truth annotations such as RGB, depth, semantic, instance, and amodal segmentation masks, bounding boxes, and detailed nutritional information per food item. We demonstrate the diversity of NV-Synth across foods, compositions, viewpoints, and lighting. As the largest open-source synthetic food dataset, NV-Synth highlights the value of physics-based simulations for enabling scalable and controllable generation of diverse photorealistic meal images to overcome data limitations and drive advancements in automated dietary assessment using computer vision. In addition to the dataset, the source code for our data generation framework is also made publicly available at https://saeejithnair.github.io/nvsynth.
Food image segmentation is an important task that has ubiquitous applications, such as estimating the nutritional value of a plate of food. Although machine learning models have been used for segmentation in this domain, food images pose several challenges. One challenge is that food items can overlap and mix, making them difficult to distinguish. Another challenge is the degree of inter-class similarity and intra-class variability, which is caused by the varying preparation methods and dishes a food item may be served in. Additionally, class imbalance is an inevitable issue in food datasets. To address these issues, two models are trained and compared, one based on convolutional neural networks and the other on Bidirectional Encoder representation for Image Transformers (BEiT). The models are trained and valuated using the FoodSeg103 dataset, which is identified as a robust benchmark for food image segmentation. The BEiT model outperforms the previous state-of-the-art model by achieving a mean intersection over union of 49.4 on FoodSeg103. This study provides insights into transfering knowledge using convolution and Transformer-based approaches in the food image domain.
Background: People's health depends on the use of proper diet as an important factor. Today, with the increasing mechanization of people's lives, proper eating habits and behaviors are neglected. On the other hand, food recommendations in the field of health have also tried to deal with this issue. But with the introduction of the Western nutrition style and the advancement of Western chemical medicine, many issues have emerged in the field of disease treatment and nutrition. Recent advances in technology and the use of artificial intelligence methods in information systems have led to the creation of recommender systems in order to improve people's health. Methods: A hybrid recommender system including, collaborative filtering, content-based, and knowledge-based models was used. Machine learning models such as Decision Tree, k-Nearest Neighbors (kNN), AdaBoost, and Bagging were investigated in the field of food recommender systems on 2519 students in the nutrition management system of a university. Student information including profile information for basal metabolic rate, student reservation records, and selected diet type is received online. Among the 15 features collected and after consulting nutrition experts, the most effective features are selected through feature engineering. Using machine learning models based on energy indicators and food selection history by students, food from the university menu is recommended to students. Results: The AdaBoost model has the highest performance in terms of accuracy with a rate of 73.70 percent. Conclusion: Considering the importance of diet in people's health, recommender systems are effective in obtaining useful information from a huge amount of data. Keywords: Recommender system, Food behavior and habits, Machine learning, Classification