Food Images Research Articles

RECIPE GENERATION BASED ON FOOD IMAGE USING MACHINE LANGUAGE

Recipe generation based on food images using machine learning (ML) is an emerging area that combines computer vision and natural language processing techniques to automatically generate recipes from images of dishes. This paper proposes a novel approach to address this problem by developing a deep learning-based system capable of recognizing ingredients and cooking methods from food images and generating coherent recipe texts. The system aims to assist home cooks, culinary enthusiasts, food bloggers, and businesses in discovering and creating new recipes from visual cues. KEYWORDS: Recipe generation, food image recognition, machine learning, computer vision, natural language processing, Recipe generation model, Cooking automation, Recipe inference, Ingredient detection, Recipe understanding, Computer vision for cooking, Recipe composition, Culinary content generation

Prediction of Probable Allergens in Food Items Using Convolutional Neural Networks

Food monitoring and nutritional analysis play a crucial role in addressing allergen-related health issues, and their importancecontinues to grow in our daily lives. In this study, we utilizeda convolutional neural network (CNN) to recognize and analyze food images, assess the nutritional content of dishes, and provide information on potential allergens. Identifying food items from images poses a significant challenge due to the wide variety of foods available. To address this, we leveraged the Logmeal API, which utilizes CNN to identify various types of meals, their ingredients, and potential allergens.

Food-related inhibitory control deficits in young male adults with obesity: Behavioral and ERP evidence from a food-related go/no-go task

PurposeObesity poses a pervasive challenge to global public health, which is linked to adverse physical health outcomes and cognitive decline. Cognitive function, particularly food-related cognitive function, plays a critical role in sustaining a healthy weight and mitigating the progression of obesity. The aim of this study was to investigate the behavioral and neuroelectronic aspects of food-related inhibitory functions in young adult males with obesity. MethodsForty-nine participants with obesity and healthy-weight were recruited (BMI = 35.83 ± 5.06 kg/m2 vs. 22.55 ± 1.73 kg/m2, age = 24.23 ± 4.55 years vs. 26.00 ± 3.97 years). A food-related Go/No-go task which included 6 distinct blocks in a randomized order was conducted to investigate the general and food-related inhibitory control. 180 stimulus images from the Food Picture Database encompassing high-calorie food, low-calorie food, and neutral images were selected. Behavioral (Go RT, Go ACC, No-go ACC) and event-related potential measures (N2 and P3 amplitude) during the food-related Go/No-go task were measured. ResultsThe main findings indicated that the group with obesity exhibited lower No-go accuracy, slower go reaction times, and smaller P3 amplitudes in high-calorie, low-calorie foods, and neutral picture, compared to the normal-weight group, but with no group difference in N2. Additionally, high-calorie food induced larger N2 and P3 amplitude than the neutral stimuli. ConclusionsYoung male adults with obesity exhibit poorer inhibitory control in both food and non-food domains, specifically in slower reaction time and reduced accuracy, featuring difficulties in neural resource recruitment during the inhibitory control process. Additionally, the P3 component could serve as sensitive indicators to reveal the neural mechanisms of inhibitory control deficits in obesity, while the N2 and P3 components may differentiate the neural differences between high-calorie foods and non-foods in inhibitory control processing. Food, especially high-calorie food, induces more neural resources and may exacerbate the poor inhibitory ability towards food in obesity. Targeted interventions such as exercise interventions, cognitive training as well as neuromodulation interventions are warranted in the future to improve impaired general and food-related inhibitory functions in the obese populations, offering both theoretical and practical frameworks for obesity prevention and treatment.

A modified time adaptive self-organizing map with stochastic gradient descent optimizer for automated food recognition system

Numerous decades of study have been devoted to associating artificial intelligence and culinary type recognition. Automated food identification systems are significant in many disciplines, comprising dietary valuation, menu analysis, and nutritional tracking. In the past, traditional image analysis algorithms caused in poor classification accuracy, but deep learning methods have enabled the identification of food types and its constituents. This study proposed a novel method to develop food recognition competence and accuracy by connecting a Stochastic Gradient Descent (SGD) optimizer to a Modified Time Adaptive Self-Organizing Map (MTA-SOM). Food arrival differences subsequent from lighting, changing perspectives, and occlusions sometimes provide challenges to traditional food recognition algorithms. In this research, propose an MTA-SOM that learns and adapts to changing food item appearances by dynamically changing its topology over time. This research leverages the self-organizing possessions of SOMs and the fine-tuning properties of SGD by relating the MTA-SOM and the SGD optimizer, thereby maximizing the advantages of both techniques. The research method includes collecting a large number of food images from a difference of cuisines and presentation styles in order to assess the effectiveness of the proposed method. This proposed method performs an extensive test and connect MTA-SOM and SGD to present approaches of food recognition. Important advances in precision and robustness are produced as the system learns to recognize food items more precisely and adapts to changes in food appearance. By automating food detection with high precision and adaptability, our method could revolutionize our capability to interact with food-related data and offer important insights into dietary practices and nutritious decisions.

Leveraging automatic personalised nutrition: food image recognition benchmark and dataset based on nutrition taxonomy

Leveraging automatic personalised nutrition: food image recognition benchmark and dataset based on nutrition taxonomy

Images Connect Us Together: Navigating a COVID-19 Local Outbreak in China Through Social Media Images

Social media images, curated or casual, have become a crucial component of communicating situational information and emotions during health crises. Despite its prevalence and significance in informational dissemination and emotional connection, there lacks a comprehensive understanding of visual crisis communication in the aftermath of a pandemic which is characterized by uncertain local situations and emotional fatigue. To fill this gap, this work collected 345,423 crisis-related posts and 65,376 original images during the Xi'an COVID-19 local outbreak in China, and adopted a mixed-methods approach to understanding themes, goals, and strategies of crisis imagery. Image clustering captured the diversity of visual themes during the outbreak, such as text images embedding authoritative guidelines and "visual diaries" recording and sharing the quarantine life. Through text classification of the post that visuals were situated in, we found that different visual themes highly correlated with the informational and emotional goals of the post text, such as adopting text images to convey the latest policies and sharing food images to express anxiety. We further unpacked nuanced strategies of crisis image use through inductive coding, such as signifying authority and triggering empathy. We discuss the opportunities and challenges of crisis imagery and provide design implications to facilitate effective visual crisis communication.

Food Image Detection Using CNN Algorithm

In the coming years, the fusion of deep learning with computer vision will usher in significant advancements in various domains, including food image analysis. A pioneering approach will be introduced to meet the rapidly increasing demand for automated recipe generation from food images, leveraging state-of-the-art deep learning techniques. The proliferation of social media platforms and the omnipresence of food-related content will necessitate efficient algorithms capable of understanding food images and generating coherent recipes. The proposed framework will integrate deep learning neural networks for robust food image recognition and sequence-to-sequence models for recipe generation. Initially, a pre-trained model will be employed on extensive food image datasets to extract pertinent features, followed by fine-tuning on domain-specific datasets to enhance accuracy. This pre-training step will enable the network to learn rich representations of various food items and their compositions. Subsequently, a sequence-to-sequence model, such as an encoder-decoder architecture equipped with attention mechanisms, will be employed to map the extracted image features to textual recipes. This model will learn to generate coherent and contextually relevant recipes from input food images, capturing the intricate details of different cuisines and dishes. Key Words: Deep Learning, Food Image Analysis, Recipe Generation, Convolutional Neural Networks.

Automated Artificial Intelligence–Based Thai Food Dietary Assessment System: Development and Validation

BackgroundDietary assessment is a fundamental component of nutrition research and plays a pivotal role in managing chronic diseases. Traditional dietary assessment methods, particularly in the context of Thai cuisine, often require extensive training and may lead to estimation errors. ObjectivesTo address these challenges, Institute of Nutrition, Mahidol University (INMU) iFood, an innovative artificial intelligence–based Thai food dietary assessment system, allows for estimating the nutritive values of dishes from food images. MethodsINMU iFood leverages state-of-the-art technology and integrates a validated automated Thai food analysis system. Users can use 3 distinct input methods: food image recognition, manual input, and a convenient barcode scanner. This versatility simplifies the tracking of dietary intake while maximizing data quality at the individual level. The core improvement in INMU iFood can be attributed to 2 key factors, namely, the replacement of Yolov4-tiny with Yolov7 and the expansion of noncarbohydrate source foods in the training image data set. ResultsThis combination significantly enhances the system’s ability to identify food items, especially in scenarios with closely packed food images, thus improving accuracy. Validation results showcase the superior performance of the INMU iFood integrated V7-based system over its predecessor, V4-based, with notable improvements in protein and fat estimation. Furthermore, INMU iFood addresses limitations by offering users the option to import additional food products via a barcode scanner, thus providing access to a vast database of nutritional information through Open Food Facts. This integration ensures users can track their dietary intake effectively, with expanded access to over 3000 food items added to or updated in the Open Food Facts database covering a wide variety of dietary choices. ConclusionsINMU iFood is a promising tool for researchers, health care professionals, and individuals seeking to monitor their dietary intake within the context of Thai cuisine and for ultimately promoting better health outcomes and facilitating nutrition-related research.

Enhancing Food Type Recognition: A Comprehensive Study on Sequential Convolutional Neural Networks for Image Classification Accuracy

Addressing the challenge of food recognition, this study investigates the effectiveness of sequential convolutional neural networks (CNNs) and their application in accurately identifying food items within images. The research introduces a novel CNN architecture, termed "sequential_2," tailored for food classification, achieving an accuracy of 89.84% on the Food Images (Food-101) dataset. Insights from the model's architecture, performance, and findings are discussed, emphasizing its potential in image classification tasks, particularly in the context of food recognition. This innovative approach aims to automate traditionally challenging and resource-intensive tasks associated with determining food attributes, creating taxonomies, and extracting nutrient information. The results highlight the potential of combining cutting-edge deep learning techniques with practical applications, showcasing a paradigm shift in the way we approach and automate the understanding of food through technology.

Padang Cuisine Classification using Deep Convolutional Neural Networks (CNN) and Transfer Learning

Padang cuisine, as an integral part of Indonesia's culinary wealth, challenges researchers to explore its potential through technological approaches. This article explores the application of Deep Convolutional Neural Networks (DCNN) and Transfer Learning for the classification of Padang cuisine. DCNN is used to understand and recognize unique visual patterns in food images, while Transfer Learning leverages existing knowledge to improve the model's performance in the specific task of classifying Padang cuisine. This process involves collecting a representative dataset, data preprocessing, and model training to create a system capable of accurately identifying various Padang dishes. Experimental results show that this approach achieves an accuracy rate of 92%, demonstrating its effectiveness in classifying different types of Padang cuisine. By engaging advanced technology, this article contributes to a deeper understanding of the integration between traditional culinary richness and innovation in the modern world of artificial intelligence. These findings indicate that AI implementation can be used to document and celebrate diverse culinary heritage, as well as support the culinary industry in managing and identifying dishes in real-time.

The associations between attentional bias to food cues, parent-report appetitive traits, and concurrent adiposity among adolescents

ObjectiveTo assess whether attentional bias to food cues and appetitive traits are independently and interactively associated with adiposity in adolescents. MethodEighty-five adolescents, 14–17-years had their attentional bias to food images measured in a sated state by computing eye tracking measures of attention (first fixation duration, cumulative fixation duration) to food and control distractor images that bordered a computer game. Parents reported adolescent appetitive traits including the food approach domains of enjoyment of food, food responsiveness, emotional overeating, and the food avoidance domains of satiety responsiveness and emotional overeating through the Children's Eating Behavior Questionnaire. ResultsFirst fixation bias to food cues was positively associated with enjoyment of food, and negatively associated with satiety responsiveness. In a series of regression models adjusted for relevant covariates, first fixation bias to food cues (β = 0.83, p = 0.007), higher food responsiveness (β = 0.74, p < 0.001), higher emotional overeating (β = 0.51, p = 0.002), and a composite appetite score (β = 1.42, p < 0.001) were each significantly associated with greater BMI z-scores. In models assessing the interactive effects between attentional bias and appetitive traits, higher first fixation bias to food cues interacted synergistically with food responsiveness and emotional overeating in relation to BMI z-score. A synergistic interaction between first fixation bias to food cues and the composite appetite score in relation to BMI z-score was also observed. ConclusionIndividuals with high attentional bias to food cues and obesogenic appetitive traits may be particularly susceptible to weight gain.

Developing a restaurant recommended system via the Vietnamese food image classification

A recommendation system is a system that recommends products and services to users based on daily online searching habits. The recommender system is applied in many fields such as job searching, health care, education, music, and tourism. However, few studies have combined computer vision and collaborative filtering to build a restaurant recommendation system in the tourism sector. In this study, we presented a solution to build a restaurant recommendation system through Vietnamese food image classification. First, we used ResNet-34 which is a variant of the convolutional neural network to classify Vietnamese food images. Then, the system applied the alternative least square technique in matrix factorization and Apache Spark in distributed computing to train the restaurant location dataset. The output was the most relevant restaurant places list to show many choices to users. The experimental datasets included the Vietnamese image and the restaurant location datasets that were collected from kaggle.com and foody.vn websites. For image classification task evaluation, we compared ResNet-34 to variants of ResNet. For the restaurant recommendation task evaluation, we compared alternative least squares with k-nearest neighbor. The comparison results show that the proposed solution is better than traditional popular models.

Patterns of affective images of animal-sourced food in Norway: Land versus sea

To get an understanding of drivers of animal-sourced protein consumption, we explored laypeople’s affective images of animal-sourced food. A national representative sample of the Norwegian population (N = 783) provided free associations to six food products originating either from livestock, capture fishery, aquaculture, or hunting. Subsequently, participants evaluated their own free associations as either positive, negative, or neutral. We found that people show different associative patterns for animal-sourced food from land than from sea. Livestock and hunting are mostly related to traditions and food, whereas capture fishery relates to production, consequences, and evaluations. People reported to have little knowledge about food products in the aquaculture category. Livestock was the most positively evaluated category, followed by hunting and capture fishery; aquaculture elicited the most negative associations. The current findings suggest a need to consider different strategies to encourage consumption of specific categories of food products.

The power of visual representations and the growing aestheticization of food. An interdisciplinary analysis of Airbnb visual content in Siena

The paper aims at exploring the role played by visual content in the processes of place touristification and commodification that are initiated and accelerated by online platforms. The analysis addresses the photographs through which Airbnb hosts illustrate and promote their accommodations, focusing on food images as the most fertile terrain for the inquiry, by virtue of the density of its aesthetic, social, geographical, and cultural meanings, constantly evoked by the photos. The analysis is based on a corpus of 250 listings located in the Municipality of Siena (Tuscany, Italy) present on Airbnb between November 2020 and February 2022, and 3960 photographs of the interiors uploaded by hosts. To this end, the authors develop an interdisciplinary method of investigation that synergistically integrates the geographical approach with interpretative models applied in the fields of semiotics and visual studies. The city of Siena is itself an excellent case study. Thanks to its small size, it was possible to map the entire set of accommodations present in the territory during the indicated period, which makes the corpus homogeneous and exhaustive; as a major tourist attraction in Tuscany and nationwide, the link between food, tourism and local identity is particularly strong and therefore worthy of study. Based on the outcomes, the conclusions show the effects and dynamics of a process of image normalization that differs profoundly from the promise of authentic and unique tourist experiences summed up in the platform's slogan “living like a local”.

MyPlate: A Diet Monitoring and Recommender Application

Introduction: Diet monitoring is of great interest in practicing a healthy diet to ensure good health by keeping track of the type and amount of food consumed. The food recommender system could provide personalized suggestions about food choices that best suit user preferences, making diet monitoring a valuable tool for maintaining good health, preventing diseases, and achieving a balanced lifestyle. Methods: This article presents a diet monitoring and recommender Android mobile application utilizing a Mask Region-based Convolutional Neural Network (Mask-RCNN) and portion analysis with a content-based filtering recommender algorithm. The application, named MyPlate, will provide a daily calorie limit and nutrition guidelines based on the user profile record, such as current and ideal weight. At the same time, MyPlate can also perform food image recognition and portion analysis from the picture captured and estimate the total calories. Finally, MyPlate provides a personalized food intake recommendation based on the previous intake record. Results: The developed MyPlate application can achieve accurate food image recognition, with the mean average precision at 0.825. Besides that, the proposed MyPlate application can perform portion analysis and provide food recommendations based on intake history at 80% similarity. Conclusion: The proposed MyPlate Android application has shown the effectiveness of incorporating machine intelligence into a diet monitoring technique that can promote healthy eating habits in the community.

Automated Food Recognition for Nutritional Analysis in Dietary Evaluation

In today's world, computer vision has made remarkable strides, especially in the realm of recognizing food images. Deep neural networks (DNN) have become a standout choice among various machine learning algorithms for their effectiveness in precisely identifying different food items captured in images. Despite the widespread adoption of DNN-based classification algorithms for food recognition, there are persistent challenges in accurately pinpointing foods due to variations in size, shape, and other defining characteristics. This paper aims to provide a brief overview of how deep learning (DL) is leveraged in food recognition and explores its wide-ranging applications. Additionally, it delves into the utilization of machine learning (ML) to construct a robust model for identifying food in images, using the Fruits and Vegetables Image Recognition Dataset. The results highlight a significant improvement in the accuracy of the proposed model, showcasing an impressive increase of around 95%.

Large-scale Food Image Datasets and Food Recognition Systems: A Survey

With the development of the Internet, more and more people like to share their daily food, which makes the Internet rich in food image data. Many researchers collect food images on the Internet, label them, and combine machine learning and other methods to achieve food recognition systems. These food recognition systems can quickly help users record their daily diet and analyze their eating habits. The advanced food identification system can currently accurately identify hundreds of foods. At the same time, it can explore the food nutrition content and give food reference suggestions by accessing the food nutrition database. In this paper, we collected and evaluated large-scale food image data sets and food recognition systems.

The Prevalence and Perception of Disordered Eating and Eating Disorders Among Dietetic Students, Interns, and Graduates at a University

Disordered eating (DE) and eating disorders (EDs) are behaviors indicative of unhealthy relationships between food and body image. They may require mental health interventions and have implications for dietetic practice when they occur in dietetic professionals. This was a cross-sectional study about them delivered over the internet to students, interns, and graduates of nutrition and dietetics at a university. About half (53%) of respondents indicated having a personal experience with DE or an ED. The results of this study may be useful for educators in designing higher education that supports future nutrition professionals in the achievement of healthy relationships with food and body image.

The Relationship Between Dancing and Eating Habits

This research paper is an analysis of how perceptions within the dance world affect eating habits. In order to answer these questions, a high school dance team of 36 girls completed a survey. Each girl was asked questions regarding eating and exercise habits, how often they danced, and their thoughts on their appearance. The completed surveys indicated that, although some disordered eating habits were present, they were not related to dance. This resulted in no correlation between the two. It was found that outside factors, such as social media and magazines, were the cause of the majority of disordered eating habits of the dance team members. The results also revealed that two-thirds of the dance team have tried to lose weight in order to be a better dancer. Despite prior research, dancing did not prove to be the root cause of disordered eating within this sample. Although the researcher's hypothesis was proven to be false, this research can now allow others to view other issues revolving around food and body image within the dance world.

푸드트럭, 소셜미디어, 유동하는 쿠바 음식문화

This article examines ethnic and regional food culture in the United States, focusing on the Jon Favreau film Chef, and analyzes the intertextual impact of sharing food photos in social media spaces on the practice of food consumption and production in the real world. A summary and main research findings are as follows. First, in Chef, America’s diverse ethnic and regional food culture is portrayed through cinematic images of various ingredients and urban spaces along the main characters’ journey to Miami, New Orleans, Austin, and Los Angeles. Each city’s locality and food culture are complexly expressed through unique street scenes that preserve the local history and culture, famous restaurants that have become local attractions, colorful and lively food images, sounds generated during the cooking process, and cheerful background music. Second, this film can enhance understanding of the motivation for setting up and operating food trucks in the U.S. For the protagonist, Carl Casper, Cuban food and food trucks serve as symbols and infrastructure that guarantee his professional autonomy and creativity as a chef. By analyzing its storyline and dialogue, this article examines the ways, meanings, and repercussions of food culture representation in the media by those in the restaurant industry who prepare and sell food themselves, not by food consumers. To recapitulate, Chef is valuable as educational content in that it vividly introduces America’s diverse ethnic and regional food culture, farmers’ markets, and food truck culture. This article will contribute to a better understanding of how the prevalence of digital media can affect and socially reorganize food culture by examining the establishment and operation of food trucks in the U.S. through the film and the relationship between food truck businesses and social network services.