1
|
Sheng G, Min W, Zhu X, Xu L, Sun Q, Yang Y, Wang L, Jiang S. A Lightweight Hybrid Model with Location-Preserving ViT for Efficient Food Recognition. Nutrients 2024; 16:200. [PMID: 38257093 PMCID: PMC10819383 DOI: 10.3390/nu16020200] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/07/2023] [Revised: 12/09/2023] [Accepted: 12/29/2023] [Indexed: 01/24/2024] Open
Abstract
Food-image recognition plays a pivotal role in intelligent nutrition management, and lightweight recognition methods based on deep learning are crucial for enabling mobile deployment. This capability empowers individuals to effectively manage their daily diet and nutrition using devices such as smartphones. In this study, we propose an Efficient Hybrid Food Recognition Net (EHFR-Net), a novel neural network that integrates Convolutional Neural Networks (CNN) and Vision Transformer (ViT). We find that in the context of food-image recognition tasks, while ViT demonstrates superiority in extracting global information, its approach of disregarding the initial spatial information hampers its efficacy. Therefore, we designed a ViT method termed Location-Preserving Vision Transformer (LP-ViT), which retains positional information during the global information extraction process. To ensure the lightweight nature of the model, we employ an inverted residual block on the CNN side to extract local features. Global and local features are seamlessly integrated by directly summing and concatenating the outputs from the convolutional and ViT structures, resulting in the creation of a unified Hybrid Block (HBlock) in a coherent manner. Moreover, we optimize the hierarchical layout of EHFR-Net to accommodate the unique characteristics of HBlock, effectively reducing the model size. Our extensive experiments on three well-known food image-recognition datasets demonstrate the superiority of our approach. For instance, on the ETHZ Food-101 dataset, our method achieves an outstanding recognition accuracy of 90.7%, which is 3.5% higher than the state-of-the-art ViT-based lightweight network MobileViTv2 (87.2%), which has an equivalent number of parameters and calculations.
Collapse
Affiliation(s)
- Guorui Sheng
- School of Information and Electrical Engineering, Ludong University, Yantai 264025, China; (G.S.); (X.Z.); (L.X.); (Q.S.); (L.W.)
| | - Weiqing Min
- Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China; (W.M.); (S.J.)
- School of Computer Science and Technology, University of Chinese Academy of Sciences, Beijing 100190, China
| | - Xiangyi Zhu
- School of Information and Electrical Engineering, Ludong University, Yantai 264025, China; (G.S.); (X.Z.); (L.X.); (Q.S.); (L.W.)
| | - Liang Xu
- School of Information and Electrical Engineering, Ludong University, Yantai 264025, China; (G.S.); (X.Z.); (L.X.); (Q.S.); (L.W.)
| | - Qingshuo Sun
- School of Information and Electrical Engineering, Ludong University, Yantai 264025, China; (G.S.); (X.Z.); (L.X.); (Q.S.); (L.W.)
| | - Yancun Yang
- School of Information and Electrical Engineering, Ludong University, Yantai 264025, China; (G.S.); (X.Z.); (L.X.); (Q.S.); (L.W.)
| | - Lili Wang
- School of Information and Electrical Engineering, Ludong University, Yantai 264025, China; (G.S.); (X.Z.); (L.X.); (Q.S.); (L.W.)
| | - Shuqiang Jiang
- Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China; (W.M.); (S.J.)
- School of Computer Science and Technology, University of Chinese Academy of Sciences, Beijing 100190, China
| |
Collapse
|
2
|
Papathanail I, Abdur Rahman L, Brigato L, Bez NS, Vasiloglou MF, van der Horst K, Mougiakakou S. The Nutritional Content of Meal Images in Free-Living Conditions-Automatic Assessment with goFOOD TM. Nutrients 2023; 15:3835. [PMID: 37686866 PMCID: PMC10490087 DOI: 10.3390/nu15173835] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/28/2023] [Revised: 08/28/2023] [Accepted: 08/30/2023] [Indexed: 09/10/2023] Open
Abstract
A healthy diet can help to prevent or manage many important conditions and diseases, particularly obesity, malnutrition, and diabetes. Recent advancements in artificial intelligence and smartphone technologies have enabled applications to conduct automatic nutritional assessment from meal images, providing a convenient, efficient, and accurate method for continuous diet evaluation. We now extend the goFOODTM automatic system to perform food segmentation, recognition, volume, as well as calorie and macro-nutrient estimation from single images that are captured by a smartphone. In order to assess our system's performance, we conducted a feasibility study with 50 participants from Switzerland. We recorded their meals for one day and then dietitians carried out a 24 h recall. We retrospectively analysed the collected images to assess the nutritional content of the meals. By comparing our results with the dietitians' estimations, we demonstrated that the newly introduced system has comparable energy and macronutrient estimation performance with the previous method; however, it only requires a single image instead of two. The system can be applied in a real-life scenarios, and it can be easily used to assess dietary intake. This system could help individuals gain a better understanding of their dietary consumption. Additionally, it could serve as a valuable resource for dietitians, and could contribute to nutritional research.
Collapse
Affiliation(s)
- Ioannis Papathanail
- ARTORG Center for Biomedical Engineering Research, University of Bern, 3008 Bern, Switzerland; (I.P.); (L.A.R.); (L.B.); (M.F.V.)
| | - Lubnaa Abdur Rahman
- ARTORG Center for Biomedical Engineering Research, University of Bern, 3008 Bern, Switzerland; (I.P.); (L.A.R.); (L.B.); (M.F.V.)
| | - Lorenzo Brigato
- ARTORG Center for Biomedical Engineering Research, University of Bern, 3008 Bern, Switzerland; (I.P.); (L.A.R.); (L.B.); (M.F.V.)
| | - Natalie S. Bez
- School of Health Professions, Bern University of Applied Sciences, 3008 Bern, Switzerland; (N.S.B.); (K.v.d.H.)
| | - Maria F. Vasiloglou
- ARTORG Center for Biomedical Engineering Research, University of Bern, 3008 Bern, Switzerland; (I.P.); (L.A.R.); (L.B.); (M.F.V.)
| | - Klazine van der Horst
- School of Health Professions, Bern University of Applied Sciences, 3008 Bern, Switzerland; (N.S.B.); (K.v.d.H.)
| | - Stavroula Mougiakakou
- ARTORG Center for Biomedical Engineering Research, University of Bern, 3008 Bern, Switzerland; (I.P.); (L.A.R.); (L.B.); (M.F.V.)
| |
Collapse
|
3
|
Huang J, Yeung AM, DuBord AY, Wolpert H, Jacobs PG, Lee WA, Drincic A, Spanakis EK, Sherr JL, Prahalad P, Fleming A, Hsiao VC, Kompala T, Lal RA, Fayfman M, Ginsberg BH, Galindo RJ, Stuhr A, Chase JG, Najafi B, Masharani U, Seley JJ, Klonoff DC. Diabetes Technology Meeting 2022. J Diabetes Sci Technol 2023; 17:1085-1120. [PMID: 36704821 PMCID: PMC10347991 DOI: 10.1177/19322968221148743] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 01/28/2023]
Abstract
Diabetes Technology Society hosted its annual Diabetes Technology Meeting from November 3 to November 5, 2022. Meeting topics included (1) the measurement of glucose, insulin, and ketones; (2) virtual diabetes care; (3) metrics for managing diabetes and predicting outcomes; (4) integration of continuous glucose monitor data into the electronic health record; (5) regulation of diabetes technology; (6) digital health to nudge behavior; (7) estimating carbohydrates; (8) fully automated insulin delivery systems; (9) hypoglycemia; (10) novel insulins; (11) insulin delivery; (12) on-body sensors; (13) continuous glucose monitoring; (14) diabetic foot ulcers; (15) the environmental impact of diabetes technology; and (16) spinal cord stimulation for painful diabetic neuropathy. A live demonstration of a device that can allow for the recycling of used insulin pens was also presented.
Collapse
Affiliation(s)
| | | | | | | | - Peter G. Jacobs
- Department of Biomedical Engineering, Oregon Health & Science University, Portland, OR, USA
| | - Wei-An Lee
- Los Angeles County+University of Southern California Medical Center, Los Angeles, CA, USA
| | | | - Elias K. Spanakis
- Baltimore Veterans Affairs Medical Center, Baltimore, MD, USA
- Division of Endocrinology, University of Maryland School of Medicine, Baltimore, MD, USA
| | | | | | | | | | - Tejaswi Kompala
- University of California, San Francisco, San Francisco, CA, USA
- Teladoc Health, Purchase, NY, USA
| | | | - Maya Fayfman
- Emory University School of Medicine, Atlanta, GA, USA
| | | | | | | | | | | | - Umesh Masharani
- University of California, San Francisco, San Francisco, CA, USA
| | | | - David C. Klonoff
- Diabetes Technology Society, Burlingame, CA, USA
- Diabetes Research Institute, Mills-Peninsula Medical Center, San Mateo, CA, USA
| |
Collapse
|
4
|
Haddad J, Vasiloglou MF, Scheidegger-Balmer F, Fiedler U, van der Horst K. Home-based cooking intervention with a smartphone app to improve eating behaviors in children aged 7-9 years: a feasibility study. DISCOVER SOCIAL SCIENCE AND HEALTH 2023; 3:13. [PMID: 37275348 PMCID: PMC10233529 DOI: 10.1007/s44155-023-00042-4] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 03/03/2023] [Accepted: 05/18/2023] [Indexed: 06/07/2023]
Abstract
Objective To develop and evaluate the feasibility of a mobile application in Swiss households and assess its impact on dietary behavior and food acceptability between children who cooked with limited parental support (intervention group) with children who were not involved in cooking (control group). Methods A ten-week randomized controlled trial was conducted online in 2020. Parents were given access to a mobile-app with ten recipes. Each recipe emphasized one of two generally disliked foods (Brussels sprouts or whole-meal pasta). Parents photographed and weighed the food components from the child's plate and reported whether their child liked the meal and target food. The main outcome measures were target food intake and acceptability analyzed through descriptive analysis for pre-post changes. Results Of 24 parents who completed the baseline questionnaires, 18 parents and their children (median age: 8 years) completed the evaluation phase. Mean child baseline Brussel sprouts and whole-meal pasta intakes were 19.0 ± 24.2 g and 86.0 ± 69.7 g per meal, respectively. No meaningful differences in intake were found post-intervention or between groups. More children reported a neutral or positive liking towards the whole-meal pasta in the intervention group compared to those in the control group. No change was found for liking of Brussel sprouts. Conclusions for practice The intervention was found to be feasible however more studies on larger samples are needed to validate feasibility. Integrating digital interventions in the home and promoting meal preparation may improve child reported acceptance of some healthy foods. Using such technology may save time for parents and engage families in consuming healthier meals.
Collapse
Affiliation(s)
- Joyce Haddad
- Bern University of Applied Sciences, School of Health Professions, Nutrition and Dietetics, Murtenstrasse 10, 3008 Bern, Switzerland
| | - Maria F. Vasiloglou
- AI in Health and Nutrition Laboratory, ARTORG Center for Biomedical Engineering Research, University of Bern, Bern, Switzerland
| | - Franziska Scheidegger-Balmer
- Bern University of Applied Sciences, School of Health Professions, Nutrition and Dietetics, Murtenstrasse 10, 3008 Bern, Switzerland
| | - Ulrich Fiedler
- Institute ICE, School of Engineering and Computer Science, Bern University of Applied Sciences, Biel/Bienne, Switzerland
| | - Klazine van der Horst
- Bern University of Applied Sciences, School of Health Professions, Nutrition and Dietetics, Murtenstrasse 10, 3008 Bern, Switzerland
| |
Collapse
|
5
|
A feasibility study to assess Mediterranean Diet adherence using an AI-powered system. Sci Rep 2022; 12:17008. [PMID: 36220998 PMCID: PMC9554192 DOI: 10.1038/s41598-022-21421-y] [Citation(s) in RCA: 4] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/22/2022] [Accepted: 09/27/2022] [Indexed: 12/29/2022] Open
Abstract
Mediterranean diet (MD) can play a major role in decreasing the risks of non-communicable diseases and preventing overweight and obesity. In order for a person to follow the MD and assess their adherence to it, proper dietary assessment methods are required. We have developed an Artificial Intelligence-powered system that recognizes the food and drink items from a single meal photo and estimates their respective serving size, and integrated it into a smartphone application that automatically calculates MD adherence score and outputs a weekly feedback report. We compared the MD adherence score of four users as calculated by the system versus an expert dietitian, and the mean difference was 3.5% and statistically not significant. Afterwards, we conducted a feasibility study with 24 participants, to evaluate the system's performance and to gather the users' and dietitians' feedback. The image recognition system achieved 61.8% mean Average Precision for the testing set and 57.3% for the feasibility study images (where the ground truth was taken as the participants' annotations). The feedback from the participants of the feasibility study was also very positive.
Collapse
|