Validation of Mobile Artificial Intelligence Technology-Assisted Dietary Assessment Tool Against Weighed Records and 24-Hour Recall in Adolescent Females in Ghana.

BACKGROUND: Important gaps exist in the dietary intake of adolescents in low- and middle-income countries (LMICs), partly due to expensive assessment methods and inaccuracy in portion-size estimation. Dietary assessment tools leveraging mobile technologies exist but only a few have been validated in LMICs. OBJECTIVE: We validated Food Recognition Assistance and Nudging Insights (FRANI), a mobile artificial intelligence (AI) dietary assessment application in adolescent females aged 12-18 y (n = 36) in Ghana, against weighed records (WR), and multipass 24-hour recalls (24HR). METHODS: Dietary intake was assessed during 3 nonconsecutive days using FRANI, WRs, and 24HRs. Equivalence of nutrient intake was tested using mixed-effect models adjusted for repeated measures, by comparing ratios (FRANI/WR and 24HR/WR) with equivalence margins at 10%, 15%, and 20% error bounds. Agreement between methods was assessed using the concordance correlation coefficient (CCC). RESULTS: Equivalence for FRANI and WR was determined at the 10% bound for energy intake, 15% for 5 nutrients (iron, zinc, folate, niacin, and vitamin B6), and 20% for protein, calcium, riboflavin, and thiamine intakes. Comparisons between 24HR and WR estimated equivalence at the 20% bound for energy, carbohydrate, fiber, calcium, thiamine, and vitamin A intakes. The CCCs by nutrient between FRANI and WR ranged between 0.30 and 0.68, which was similar for CCC between 24HR and WR (ranging between 0.38 and 0.67). Comparisons of food consumption episodes from FRANI and WR found 31% omission and 16% intrusion errors. Omission and intrusion errors were lower when comparing 24HR with WR (21% and 13%, respectively). CONCLUSIONS: FRANI AI-assisted dietary assessment could accurately estimate nutrient intake in adolescent females compared with WR in urban Ghana. FRANI estimates were at least as accurate as those provided through 24HR. Further improvements in food recognition and portion estimation in FRANI could reduce errors and improve overall nutrient intake estimations.

Feasibility of Using an Artificial Intelligence-based Telephone Application for Dietary Assessment and Nudging to Improve the Quality of Food Choices of Female Adolescents in Vietnam: Evidence from a Randomized Pilot Study.

Background: Adolescent nutrition has faced a policy neglect, partly owing to the gaps in dietary intake data for this age group. The Food Recognition Assistance and Nudging Insights (FRANI) is a smartphone application validated for dietary assessment and to influence users toward healthy food choices. Objectives: This study aimed to assess the feasibility (adherence, acceptability, and usability) of FRANI and its effects on food choices and diet quality in female adolescents in Vietnam. Methods: Adolescents (N = 36) were randomly selected from a public school and allocated into 2 groups. The control group received smartphones with a version of FRANI limited to dietary assessment, whereas the intervention received smartphones with gamified FRANI. After the first 4 wk, both groups used gamified FRANI for further 2 wk. The primary outcome was the feasibility of using FRANI as measured by adherence (the proportion of completed food records), acceptability and usability (the proportion of participants who considered FRANI acceptable and usable according to answers of a Likert questionnaire). Secondary outcomes included the percentage of meals recorded, the Minimum Dietary Diversity for Women (MDDW) and the Eat-Lancet Diet Score (ELDS). Dietary diversity is important for dietary quality, and sustainable healthy diets are important to reduce carbon emissions. Poisson regression models were used to estimate the effect of gamified FRANI on the MDDW and ELDS. Results: Adherence to the application was 82% and the percentage of meals recorded was 97%. Acceptability and usability were 97%. MDDW in the intervention group was 1.07 points (95% CI: 0.98, 1.18; P = 0.13) greater than that in the control (constant = 4.68); however, the difference was not statistically significant. Moreover, ELDS in the intervention was 1.09 (95% CI: 1.01, 1.18; P = 0.03) points greater than in the control (constant = 3.67). Conclusions: FRANI was feasible and may be effective to influence users toward healthy food choices. Research is needed for FRANI in different contexts and at scale.The trial was registered at the International Standard Randomized Controlled Trial Number as ISRCTN 10681553.

Exploring an Artificial Intelligence-Based, Gamified Phone App Prototype to Track and Improve Food Choices of Adolescent Girls in Vietnam: Acceptability, Usability, and Likeability Study.

BACKGROUND: Adolescents' consumption of healthy foods is suboptimal in low- and middle-income countries. Adolescents' fondness for games and social media and the increasing access to smartphones make apps suitable for collecting dietary data and influencing their food choices. Little is known about how adolescents use phones to track and shape their food choices. OBJECTIVE: This study aimed to examine the acceptability, usability, and likability of a mobile phone app prototype developed to collect dietary data using artificial intelligence-based image recognition of foods, provide feedback, and motivate users to make healthier food choices. The findings were used to improve the design of the app. METHODS: A total of 4 focus group discussions (n=32 girls, aged 15-17 years) were conducted in Vietnam. Qualitative data were collected and analyzed by grouping ideas into common themes based on content analysis and ground theory. RESULTS: Adolescents accepted most of the individual- and team-based dietary goals presented in the app prototype to help them make healthier food choices. They deemed the overall app wireframes, interface, and graphic design as acceptable, likable, and usable but suggested the following modifications: tailored feedback based on users' medical history, anthropometric characteristics, and fitness goals; new language on dietary goals; provision of information about each of the food group dietary goals; wider camera frame to fit the whole family food tray, as meals are shared in Vietnam; possibility of digitally separating food consumption on shared meals; and more appealing graphic design, including unique badge designs for each food group. Participants also liked the app's feedback on food choices in the form of badges, notifications, and statistics. A new version of the app was designed incorporating adolescent's feedback to improve its acceptability, usability, and likability. CONCLUSIONS: A phone app prototype designed to track food choice and help adolescent girls from low- and middle-income countries make healthier food choices was found to be acceptable, likable, and usable. Further research is needed to examine the feasibility of using this technology at scale.

Relative validity of a mobile AI-technology-assisted dietary assessment in adolescent females in Vietnam.

BACKGROUND: There is a gap in data on dietary intake of adolescents in low- and middle-income countries (LMICs). Traditional methods for dietary assessment are resource intensive and lack accuracy with regard to portion-size estimation. Technology-assisted dietary assessment tools have been proposed but few have been validated for feasibility of use in LMICs. OBJECTIVES: We assessed the relative validity of FRANI (Food Recognition Assistance and Nudging Insights), a mobile artificial intelligence (AI) application for dietary assessment in adolescent females (n = 36) aged 12-18 y in Vietnam, against a weighed records (WR) standard and compared FRANI performance with a multi-pass 24-h recall (24HR). METHODS: Dietary intake was assessed using 3 methods: FRANI, WR, and 24HRs undertaken on 3 nonconsecutive days. Equivalence of nutrient intakes was tested using mixed-effects models adjusting for repeated measures, using 10%, 15%, and 20% bounds. The concordance correlation coefficient (CCC) was used to assess the agreement between methods. Sources of errors were identified for memory and portion-size estimation bias. RESULTS: Equivalence between the FRANI app and WR was determined at the 10% bound for energy, protein, and fat and 4 nutrients (iron, riboflavin, vitamin B-6, and zinc), and at 15% and 20% bounds for carbohydrate, calcium, vitamin C, thiamin, niacin, and folate. Similar results were observed for differences between 24HRs and WR with a 20% equivalent bound for all nutrients except for vitamin A. The CCCs between FRANI and WR (0.60, 0.81) were slightly lower between 24HRs and WR (0.70, 0.89) for energy and most nutrients. Memory error (food omissions or intrusions) was ∼21%, with no clear pattern apparent on portion-size estimation bias for foods. CONCLUSIONS: AI-assisted dietary assessment and 24HRs accurately estimate nutrient intake in adolescent females when compared with WR. Errors could be reduced with further improvements in AI-assisted food recognition and portion estimation.

Measuring adherence, acceptability and likability of an artificial-intelligence-based, gamified phone application to improve the quality of dietary choices of adolescents in Ghana and Vietnam: Protocol of a randomized controlled pilot test.

Unhealthy diets are a critical global concern while dietary measure methods are time consuming and expensive. There is limited evidence that phone-based interventions can improve nutrition data collection and dietary quality, especially for adolescents in developing countries. We developed an artificial-intelligence-based phone application called Food Recognition Assistance and Nudging Insights (FRANI) to address these problems. FRANI can recognize foods in images, track food consumption, display statistics and use gamified nudges to give positive feedback on healthy food choice. This study protocol describes the design of new pilot studies aimed at measuring the feasibility (acceptability, adherence, and usability) of FRANI and its effects on the quality of food choice of adolescents in Ghana and Vietnam. In each country, 36 adolescents (12-18 years) will be randomly allocated into two groups: The intervention group with the full version of FRANI and the control group with the functionality limited to image recognition and dietary assessment. Participants in both groups will have their food choices tracked for four weeks. The control groups will then switch to the full version of FRANI and both groups will be tracked for a further 2 weeks to assess acceptability, adherence, and usability. Analysis of outcomes will be by intent to treat and differences in outcomes between intervention and control group will use Poisson and odds ratio regression models, accounting for repeated measures at individual levels. If deemed feasible, acceptable and usable, FRANI will address gaps in the literature and advance the nutrition field by potentially improving the quality of food choices of adolescent girls in developing countries. This pilot study will also provide insights on the design of a large randomized controlled trial. The functioning and dissemination of FRANI can be an important step towards highly scalable nutrition data collection and healthier food choices for a population at risk of malnutrition. The study protocol and the methods and materials were approved by the Institutional Review Board (IRB) of the IFPRI on April 29th, 2020 (registration number #00007490), the Thai Nguyen National Hospital on April 14th, 2020 (protocol code 274/DDD-BVTWTN) and the University of Ghana on August 10th, 2020 (Federalwide Assurance FWA 00001824; NMIMR-IRB CPN 078-19/20). The study protocol was registered in the International Standard Randomized Controlled Trial Number (ISRCTN 10681553; https://doi.org/10.1186/ISRCTN10681553) on November 12, 2021.

Accuracy of a Smartphone-Based Object Detection Model, PlantVillage Nuru, in Identifying the Foliar Symptoms of the Viral Diseases of Cassava-CMD and CBSD.

Nuru is a deep learning object detection model for diagnosing plant diseases and pests developed as a public good by PlantVillage (Penn State University), FAO, IITA, CIMMYT, and others. It provides a simple, inexpensive and robust means of conducting in-field diagnosis without requiring an internet connection. Diagnostic tools that do not require the internet are critical for rural settings, especially in Africa where internet penetration is very low. An investigation was conducted in East Africa to evaluate the effectiveness of Nuru as a diagnostic tool by comparing the ability of Nuru, cassava experts (researchers trained on cassava pests and diseases), agricultural extension officers and farmers to correctly identify symptoms of cassava mosaic disease (CMD), cassava brown streak disease (CBSD) and the damage caused by cassava green mites (CGM). The diagnosis capability of Nuru and that of the assessed individuals was determined by inspecting cassava plants and by using the cassava symptom recognition assessment tool (CaSRAT) to score images of cassava leaves, based on the symptoms present. Nuru could diagnose symptoms of cassava diseases at a higher accuracy (65% in 2020) than the agricultural extension agents (40-58%) and farmers (18-31%). Nuru's accuracy in diagnosing cassava disease and pest symptoms, in the field, was enhanced significantly by increasing the number of leaves assessed to six leaves per plant (74-88%). Two weeks of Nuru practical use provided a slight increase in the diagnostic skill of extension workers, suggesting that a longer duration of field experience with Nuru might result in significant improvements. Overall, these findings suggest that Nuru can be an effective tool for in-field diagnosis of cassava diseases and has the potential to be a quick and cost-effective means of disseminating knowledge from researchers to agricultural extension agents and farmers, particularly on the identification of disease symptoms and their management practices.

A Mobile-Based Deep Learning Model for Cassava Disease Diagnosis.

Convolutional neural network (CNN) models have the potential to improve plant disease phenotyping where the standard approach is visual diagnostics requiring specialized training. In scenarios where a CNN is deployed on mobile devices, models are presented with new challenges due to lighting and orientation. It is essential for model assessment to be conducted in real world conditions if such models are to be reliably integrated with computer vision products for plant disease phenotyping. We train a CNN object detection model to identify foliar symptoms of diseases in cassava (Manihot esculenta Crantz). We then deploy the model in a mobile app and test its performance on mobile images and video of 720 diseased leaflets in an agricultural field in Tanzania. Within each disease category we test two levels of severity of symptoms-mild and pronounced, to assess the model performance for early detection of symptoms. In both severities we see a decrease in performance for real world images and video as measured with the F-1 score. The F-1 score dropped by 32% for pronounced symptoms in real world images (the closest data to the training data) due to a decrease in model recall. If the potential of mobile CNN models are to be realized our data suggest it is crucial to consider tuning recall in order to achieve the desired performance in real world settings. In addition, the varied performance related to different input data (image or video) is an important consideration for design in real world applications.

Deep Learning for Image-Based Cassava Disease Detection.

Cassava is the third largest source of carbohydrates for human food in the world but is vulnerable to virus diseases, which threaten to destabilize food security in sub-Saharan Africa. Novel methods of cassava disease detection are needed to support improved control which will prevent this crisis. Image recognition offers both a cost effective and scalable technology for disease detection. New deep learning models offer an avenue for this technology to be easily deployed on mobile devices. Using a dataset of cassava disease images taken in the field in Tanzania, we applied transfer learning to train a deep convolutional neural network to identify three diseases and two types of pest damage (or lack thereof). The best trained model accuracies were 98% for brown leaf spot (BLS), 96% for red mite damage (RMD), 95% for green mite damage (GMD), 98% for cassava brown streak disease (CBSD), and 96% for cassava mosaic disease (CMD). The best model achieved an overall accuracy of 93% for data not used in the training process. Our results show that the transfer learning approach for image recognition of field images offers a fast, affordable, and easily deployable strategy for digital plant disease detection.