Non-nutritive sweeteners (NNS) are frequently used as substitutes for added sugars. NNS are difficult to study due to the inability to accurately measure the amounts individuals consume, as well as limitations in study design and methods, including reliance on observational study designs, the use of diet soda as a proxy of total NNS intake, and the grouping of NNS into a single category rather than studying NNS as individual products. New dietary assessment methods and improved study design and methods are needed to advance researchers' abilities to study NNS and their impact on consumers' health. The objectives of this dissertation were to 1) determine validity and reproducibility of a novel NNS food frequency questionnaire (NNS-FFQ), 2) develop methodology for an objective NNS urinary biomarker, 3) identify an appropriate carrier for NNS intake in studies, and 4) examine the literature on the relationship between NNS and weight-related outcomes based on study design and methods. Objectives 1 and 2: participants (n=125) completed three 24-hr dietary recalls, the NNS-FFQ, and 2 24-hr urine samples. NNS intake via NNS-FFQ and recalls were compared using Bland-Altman analyses, with agreement levels ranging from 92.7-99.2% for individual NNS types and total intake. The NNS biomarker methodology was developed using ultra performance liquid chromatography (UPLC-MS/MS), which analyzes each sample for the presence of NNS and related metabolites. This method observed a range from very strong presence of NNS to not detectable, indicating that this biomarker could identify specific NNS consumption (n=9). Objective 3: a sensory evaluation (n=67) was conducted to identify if applesauce or water was a more appropriate carrier for NNS for future interventions. Applesauce was preferred (sucralose=83.6%; aspartame=79.1%; stevia=74.6%) significantly more than water (p≤0.001), indicating that applesauce could be used as an acceptable carrier of NNS in research studies. Objective 4: a systematic literature review focusing on study design and methods used in investigations on NNS and weight-related outcomes found that 81% of RCT had improved weight outcomes, while 76% of observational studies had higher weight outcomes. Improving NNS study design and methods will increase the quality of research conducted on NNS and related health outcomes. / Ph. D. / Artificial sweeteners [non-nutritive sweeteners (NNS)] are often used to replace added sugars. NNS are difficult to study due to the inability to accurately measure the amounts individuals consume, as well as limitations in study design and methods, including reliance on observational study designs, the use of diet soda to represent total NNS intake, and the grouping of NNS into a single category rather than studying NNS as individual products. New dietary assessment tools and improved study design and methods are needed to allow researchers better to study NNS and their impact on health. The objectives of this dissertation were to 1) to determine the ability of a NNS food frequency questionnaire (NNS-FFQ) to measure typical NNS intake, 2) develop methodology for an objective NNS urinary biomarker, 3) identify an appropriate carrier for NNS intake in studies, and 4) to review the currently available research on the relationship between NNS and weight status. Objectives 1 and 2: 125 participants completed three 24-hr dietary recalls, the NNS-FFQ to measure usual NNS intake, and 24-hour urine samples for a NNS biomarker. Amounts of NNS that consumers reported in recalls were compared to amounts reported in the NNS-FFQ, with the tools finding similar NNS amounts in participants’ diets. The NNS biomarker methodology was developed using ultra performance liquid chromatography (UPLC-MS/MS). This rapid method measures the presence of NNS and related products (saccharin, acesulfame potassium, sucralose, steviol glucuronide, and erythritol) in urine. Among 9 participants, this method identified wide differences, ranging from strong presence of NNS to not measureable. Objective 3: participants (n=67) completed a sensory evaluation (taste testing) study to identify a more appropriate food or beverage (applesauce or water) for NNS (sucralose, aspartame, and stevia) to be mixed with in research studies. Applesauce was preferred over water for all sweeteners (>74%), indicating that applesauce would be an acceptable NNS carrier. Objective 4: a systematic review on study design and research methods used in studies on NNS and weight-related outcomes. Improving NNS measurement tools and study design methods will improve the quality of research that can be conducted on NNS and related health outcomes.

Doctor of Philosophy / Department of Human Nutrition / Edgar Chambers IV / Because healthy food products do not have a specific tool used for consumer screening based on consumers’ diet or degree of healthy eating habits, this study aimed to determine a set of questions that could classify consumers who belong in a different status according to the Stages of Change model, including those who have a different diet quality based on their Healthy Eating Index (HEI) score. The surveys were conducted in the United States (US) and Thailand in order to determine applicability to varying countries. The Food Neophobia Scale (FNS), Food Involvement Scale (FIS), and Health and Taste Attitude Scale (HTAS) were included in the questionnaire together with a set of Stages of Change questions and a 7-day, self-administered food recall questionnaire. The HEI interpretation of US and Thai consumer scores illustrated that the majority of both belonged to the Need Improvement group. The Stages of Change model indicated most consumers thought they had healthy diets. According to FNS, FIS, and HTAS, US consumers are more involved in food activities and are more open to trying new foods or unfamiliar foods than Thais. Furthermore, consumers who belong in different groups, according to the Stages of Change model, responded differently to some HTAS subscales. However, statements from FNS, FIS, and HTAS were not capable of distinguishing consumers belonging in different groups according to HEI scores or belonging in different stages according to the Stages of Change Model. Considering all possible methods from those listed above for screening consumers, the Stages of Change model may be the best way to segment consumers interested in healthier eating. Using the Stages of Change required less time and the least effort from consumers because there were only three questions; and interpreting results does not require calculation or analysis.

Introdução: Índices dietéticos (ID) têm sido utilizados para estudar associações entre alimentação e doenças cardiovasculares (DCV). Um ID combina e resume aspectos de uma recomendação ou guia alimentar. Esses aspectos geralmente são a quantidade de nutrientes, alimentos ou grupos de alimentos a serem consumidos em um período de tempo (dia / semana / mês). A dieta do Programa Alimentar Brasileiro Cardioprotetor (DICA Br) classificou os alimentos em quatro grupos e definiu a quantidade diária a ser consumida. Uma vez que os aspectos nutricionais do DICA Br são diferentes de outras recomendações alimentares, não é possível utilizar os ID existentes para associar essa dieta com as DCV. Por isso, faz-se necessário o desenvolvimento de ID que considere as características da DICA Br. Objetivo: Descrever o desenvolvimento do ID DICA Br, avaliar a sua consistência interna, validade de constructo e as características da população associadas ao índice. Métodos: Foram utilizados dados de baseline do ensaio clínico randomizado DICA Br (www.clinicaltrials.gov; NCT01620398). Os quatro grupos alimentares da dieta estudada foram adotados como componentes do índice. Os critérios para atribuir pontos foram definidos a priori com base nos princípios e recomendações da DICA Br. Cada componente do índice recebeu pontos que variam de 0 a 10, portanto a pontuação total variou de 0 a 40. A consistência interna foi avaliada por meio dos coeficientes de correlação entre a pontuação total e de cada componente do índice, assim como pelo alpha de Cronbach. A validade de constructo foi avaliada verificando como nutrientes se associam com o índice por meio de testes para tendência linear. Análises de regressão linear bruta e ajustada foram realizadas para avaliar as características da população associadas ao índice. Resultados: A análise incluiu 2044 indivíduos (58,6 por cento homens). A média do índice total foi maior entre as mulheres. Os componentes do índice apresentaram baixas correlações entre si e as correlações entre cada componente isolado e o índice total foram >0,40. O alpha de Cronbach foi 0,66. Maiores pontuações no índice estiveram inversamente associadas com o consumo de energia, gordura total, gordura monoinsaturada, colesterol e diretamente associadas com a ingestão de carboidratos e fibras. Homens hipertensos e mulheres diabéticas apresentaram maiores pontuações, enquanto homens fumantes apresentaram menores pontuações. Conclusões: O ID DICA Br apresentou confiabilidade e validade de constructo satisfatórias, refletiu a ingestão de nutrientes chaves e detectou características dos indivíduos que se associam com a DICA Br. / Background: The diet of the Brazilian Cardioprotective Nutritional Programme (BALANCE) classified food into four groups and set the daily amount to be consumed. The nutritional aspects of BALANCE are different from other dietary recommendations, therefore it is not possible to use existing diet indexes (DI) to associate this diet with cardiovascular disease. Objective: To describe the development of BALANCE DI, evaluate its internal consistency, construct validity and population characteristics associated with the index. Methods: We analyzed baseline data from BALANCE randomized clinical trial (www.clinicaltrials.gov, NCT01620398). The four food groups of the studied diet were adopted as index components. Each index component received points ranging from 0 to 10 and the total score ranged from 0 to 40. The internal consistency was evaluated by means of correlation coefficients between total and each component index score, as well as the Cronbach´s alpha coefficient. The construct validity was assessed by checking how nutrients are associated with the index. Crude and adjusted linear regression analyses were performed to evaluate the characteristics of the population that are associated with the index. Results: The analysis included 2044 subjects (58.6 per cent men). The average of the total index was higher among women. The components of the index showed low correlations with each other and the correlations between each individual component and the total index were > 0.40. Cronbach\'s alpha coefficient was 0.66. High scores in the index were inversely associated with the intake of energy, total fat, monounsaturated fat, cholesterol and directly associated with the intake of carbohydrates and fiber. Hypertensive men and diabetic women had higher scores, while male smokers had lower scores. Conclusions: The BALANCE DI showed satisfactory reliability and construct validity, reflected the intake of key nutrients and detected characteristics of individuals that are associated with the BALANCE diet.

Thesis submitted in fulfilment of the requirements for the degree Master of Technology: Consumer Sciences: Food and Nutrition in the Faculty of Applied Sciences at the Cape Peninsula University of Technology 2013 / Child undernutrition is the outcome of numerous complex and interrelated factors. It is considered a worldwide health concern, also in Botswana. Undernourished infants are especially vulnerable to develop nutritional deficiencies and diseases. Tsabana was developed as a weaning food to improve the diet quality of undernourished infants in Botswana but also as a dietary intervention product for the four-month- to five-year-old age group to improve the nutritional status of Batswana children. The study undertook a theoretical nutritional appraisal of Tsabana and assessed its acceptance and use in rural districts of Botswana. A quantitative approach that incorporated a survey was used as the research design for the study. A pre-tested questionnaire consisting of close-ended questions in the multiple choice format was used to collect the data on the field acceptance and use of Tsabana. The questionnaire was completed by 105 caregivers of infants aged six to 36 months old visiting the selected clinics with the research sites the Ngamiland, North East, Central and Kweneng rural districts. The energy, macro- and micronutrient content adequacy of Tsabana was compared to the Codex Alimentarius Standard for Processed Cereal-based Foods for Infants and Young Children (Codex Standard 074) and the Proposed Nutrient Composition for Fortified Complementary Foods (PNCFCF). Sorghum and soya are the two major ingredients of Tsabana which contribute to its protein and energy provision. The nutrient appraisal revealed that the level of energy provided by Tsabana is more than the specified Codex Alimentarius Standard composition but slightly less than the PNCFCF specified range, whereas the protein level is within the specified ranges as put forward by both the Codex composition and the PNCFCF, respectively. The results further revealed that in comparison to the PNCFCF the content of vitamins A, E, C and pyridoxine, zinc, iodine and iron were less than the specified ranges. The content of vitamins A and D met the compositional guidelines of the Codex Alimentarius Standard. The micronutrient provision of Tsabana in relation to the World Health Organization (WHO) recommended nutrient intakes was mostly adequate except for pyridoxine which was inadequately provided for both seven- to 18-month-old (33% provision) and one- to three-year-old (52% provision) infants at the daily rations of 75 gram (g) and 200 g powder respectively. Some minerals also were provided at inadequate amounts, e.g. iodine at a daily provision of 42% as well as iron at 26% for seven- to 12-month-old and 12- to 18-month-old infants both provided at a daily ration of 75 g Tsabana powder. Zinc was only provided at 19% of the WHO recommended intake for seven- to 12-month-old and 14.8% for 12- to 18-month-old infants. As with the absence of omega-3 and omega-6 fatty acids, the minerals selenium and copper are not provided by Tsabana. The micronutrient fortification compounds used in Tsabana for the mineral and vitamin additions as vitamin A palmitate, cholecalciferol, vitamin E acetate, calcium pantothenate, carbonate and tri-calcium phosphate (ratio 1:18:482), zinc sulphate and potassium iodate are readily bioavailable, but not the vitamin A and vitamin E compounds used. The iron fortificant used is not stipulated on the Tsabana package. The field survey results revealed that Tsabana was well accepted by the infants as perceived by their caregivers as the majority indicated that Tsabana was enjoyed (80%) and that it tasted nice (71.4%). The majority (78.1%) of the caregivers also indicated that the wellbeing of the infants improved on receiving Tsabana. Most (51.4%) of them also knew that Tsabana is more nutritious than other cooked porridges. Tsabana though was not used properly as most of the caregivers fed Tsabana to the infant only once a day (60.9%) instead of two to three times (for infants six- to 18-month-olds) and did not use enough water (3½ cups) (76.2%) to cook one feeding of Tsabana as stipulated on the Tsabana package. The product use factors, which include the number of daily Tsabana feedings provided, the addition of sugar to the cooked Tsabana, the amount of water used to cook one feeding of Tsabana and the cooking period of one feeding were the factors significantly (p < 0.05) linked to the perceived infant acceptance of Tsabana. The number of daily Tsabana feedings was linked to the perceived enjoyment of Tsabana (p < 0.05) and its perceived texture (p < 0.05) and colour (p < 0.05) acceptance. While the addition of sugar to the cooked Tsabana was linked to its perceived enjoyment (p < 0.05), the amount of water used to its perceived taste acceptance (p < 0.05) and the cooking period of one feeding to its perceived colour acceptance (p < 0.05). For instance, more caregivers in relation to the enjoyment of Tsabana and its colour and texture acceptance who perceived it not to be enjoyed and not having a nice colour or texture, indicated not to feed it daily in comparison to caregivers who perceived that Tsabana was enjoyed by the infants and that it had a nice colour and texture. The caregiver perceived infant wellbeing improvement through the provision of Tsabana and the perceived difference between Tsabana and other cooked porridges were the only two clinic-related factors which significantly (p < 0.05) influenced the infant acceptance of Tsabana as perceived by the caregivers. For instance, while the majority (86.2%, 84.8% and 80%, respectively) of the caregivers who perceived the infant taste acceptance of Tsabana as very nice, nice or acceptable indicated that providing Tsabana to the infant improved the infant’s wellbeing, the majority (80%) who perceived it as not tasting nice indicated that providing Tsabana to the infant had not improved the infant’s wellbeing. More than half of the caregivers who perceived the infant texture acceptance of Tsabana as acceptable (59.4%) or nice (56.8%) indicated that Tsabana provides more nutrition to the infant. However, less than half (46.2%) of the caregivers who perceived the texture of Tsabana as not nice indicated that it provides more nutrition to the infant. The caregiver education level was the only demographic factor that significantly (p < 0.05) influenced the acceptance of Tsabana. The majority (90.3%) of the caregivers who attained a secondary education level (form 1 to form 4) and higher, perceived the infant texture acceptance of Tsabana as nice, acceptable or not nice. In contrast, the majority who perceived the texture acceptance as very nice attained either no schooling/standard one to seven (73.9%) and not a secondary level or higher (26.1%). It could be speculated that caregivers who attained a higher education level may be more critical of such attributes such as the texture acceptance of Tsabana. Most (59%) of the caregivers indicated that they did not receive any information regarding Tsabana from the clinic personnel. Caregiver education on Tsabana and infant nutrition will impart sufficient knowledge and skills to prepare, store and adequately feed Tsabana to their infants as well as provide a variety of foods in adequate amounts to contribute to their optimal growth and development. This is a vital undertaking as it was evident from the study that most of the factors that influenced the acceptance of Tsabana were the product use factors. The study additionally found that foods such as meat, poultry and fish, as well as fruit and vegetables, which should be included in the daily dietary provision, were provided to most of the study infants on a weekly basis only.

Introdução: Índices dietéticos (ID) têm sido utilizados para estudar associações entre alimentação e doenças cardiovasculares (DCV). Um ID combina e resume aspectos de uma recomendação ou guia alimentar. Esses aspectos geralmente são a quantidade de nutrientes, alimentos ou grupos de alimentos a serem consumidos em um período de tempo (dia / semana / mês). A dieta do Programa Alimentar Brasileiro Cardioprotetor (DICA Br) classificou os alimentos em quatro grupos e definiu a quantidade diária a ser consumida. Uma vez que os aspectos nutricionais do DICA Br são diferentes de outras recomendações alimentares, não é possível utilizar os ID existentes para associar essa dieta com as DCV. Por isso, faz-se necessário o desenvolvimento de ID que considere as características da DICA Br. Objetivo: Descrever o desenvolvimento do ID DICA Br, avaliar a sua consistência interna, validade de constructo e as características da população associadas ao índice. Métodos: Foram utilizados dados de baseline do ensaio clínico randomizado DICA Br (www.clinicaltrials.gov; NCT01620398). Os quatro grupos alimentares da dieta estudada foram adotados como componentes do índice. Os critérios para atribuir pontos foram definidos a priori com base nos princípios e recomendações da DICA Br. Cada componente do índice recebeu pontos que variam de 0 a 10, portanto a pontuação total variou de 0 a 40. A consistência interna foi avaliada por meio dos coeficientes de correlação entre a pontuação total e de cada componente do índice, assim como pelo alpha de Cronbach. A validade de constructo foi avaliada verificando como nutrientes se associam com o índice por meio de testes para tendência linear. Análises de regressão linear bruta e ajustada foram realizadas para avaliar as características da população associadas ao índice. Resultados: A análise incluiu 2044 indivíduos (58,6 por cento homens). A média do índice total foi maior entre as mulheres. Os componentes do índice apresentaram baixas correlações entre si e as correlações entre cada componente isolado e o índice total foram >0,40. O alpha de Cronbach foi 0,66. Maiores pontuações no índice estiveram inversamente associadas com o consumo de energia, gordura total, gordura monoinsaturada, colesterol e diretamente associadas com a ingestão de carboidratos e fibras. Homens hipertensos e mulheres diabéticas apresentaram maiores pontuações, enquanto homens fumantes apresentaram menores pontuações. Conclusões: O ID DICA Br apresentou confiabilidade e validade de constructo satisfatórias, refletiu a ingestão de nutrientes chaves e detectou características dos indivíduos que se associam com a DICA Br. / Background: The diet of the Brazilian Cardioprotective Nutritional Programme (BALANCE) classified food into four groups and set the daily amount to be consumed. The nutritional aspects of BALANCE are different from other dietary recommendations, therefore it is not possible to use existing diet indexes (DI) to associate this diet with cardiovascular disease. Objective: To describe the development of BALANCE DI, evaluate its internal consistency, construct validity and population characteristics associated with the index. Methods: We analyzed baseline data from BALANCE randomized clinical trial (www.clinicaltrials.gov, NCT01620398). The four food groups of the studied diet were adopted as index components. Each index component received points ranging from 0 to 10 and the total score ranged from 0 to 40. The internal consistency was evaluated by means of correlation coefficients between total and each component index score, as well as the Cronbach´s alpha coefficient. The construct validity was assessed by checking how nutrients are associated with the index. Crude and adjusted linear regression analyses were performed to evaluate the characteristics of the population that are associated with the index. Results: The analysis included 2044 subjects (58.6 per cent men). The average of the total index was higher among women. The components of the index showed low correlations with each other and the correlations between each individual component and the total index were > 0.40. Cronbach\'s alpha coefficient was 0.66. High scores in the index were inversely associated with the intake of energy, total fat, monounsaturated fat, cholesterol and directly associated with the intake of carbohydrates and fiber. Hypertensive men and diabetic women had higher scores, while male smokers had lower scores. Conclusions: The BALANCE DI showed satisfactory reliability and construct validity, reflected the intake of key nutrients and detected characteristics of individuals that are associated with the BALANCE diet.

<p>Dietary supplement (DS) use is highly prevalent in the U.S. and can contribute substantial amounts of micronutrients to the one-half of U.S. adults and nearly one-third of U.S. children who use them. Despite their pervasive use, the best practices for assessing the prevalence of use of and measuring nutrient intakes from supplemental sources remain largely unknown, and currently no metric designed to assess total micronutrient exposures from foods, beverages, and DS exists. Comprehensive and robust methods for assessing the prevalence of use of and nutrient intakes from DS are critical for population-level research and surveillance. </p> <p>Therefore, the overarching aims of the research presented in this dissertation were to 1) identify the most comprehensive method for assessing the prevalence of use of and estimating nutrient intakes from DS, 2) characterize DS use and examine trends in overall and micronutrient-containing DS use among the U.S. population using multiple modes of DS assessment, 3) develop the first micronutrient-based diet quality index designed to capture usual intakes from all sources of under-consumed micronutrients among the U.S. population (i.e., the Total Nutrient Index (TNI)) and 4) to determine if the TNI is a useful measure for assessing total micronutrient exposures among the U.S. adult population.</p> <p>National Health and Nutrition Examination Survey (NHANES) 2011-2014 data among U.S. adults were employed to evaluate four combinations of methods constructed from two data collection instruments (i.e., the 24-hour dietary recall (24HR) and the NHANES in-home inventory with a frequency-based DS prescription medicine questionnaire (DSMQ)) for measuring the prevalence of use of and amounts of selected nutrients from DS. The results from the analysis indicated that a higher prevalence of DS use is reported when using frequency-based methods, like the DSMQ, but higher nutrient amounts are reported on the 24HR. Thus, combining the DSMQ with at least one 24HR (i.e., DSMQ or ≥1 24HR) provides the most comprehensive approach for measuring DS use and estimating nutrient intakes from supplemental sources among U.S. adults.</p> <p>Differences in the estimated prevalence of use of DS on the DSMQ or ≥1 24HR as compared with the DSMQ only were also noted among a nationally representative sample of the U.S. population from the 2007-2018 NHANES. When characterizing DS use and examining trends in overall and micronutrient-containing DS use among the U.S. population using multiple modes of DS assessment, increases in both overall and micronutrient-containing DS use were observed over time, but the prevalence of use was lower on the DSMQ than the DSMQ or ≥1 24HR among the total population and across all population subgroups. Therefore, the findings from this study were congruent with our previous analysis and contributed to a growing body of literature that utilizes multiple methods of DS assessment to improve the comprehensiveness of DS exposure assessment. </p> <p>The insights garnered from identifying the most comprehensive method for approximating micronutrient intakes from supplemental sources were implemented in the development and initial application of the first micronutrient-based diet quality index designed to capture usual intakes from all sources of under-consumed micronutrients among the U.S. population, named the TNI. The TNI assesses U.S. adults’ total nutrient intakes relative to the Recommended Dietary Allowance or Adequate Intake for eight under-consumed micronutrients identified by the Dietary Guidelines for Americans: calcium, magnesium, potassium, choline, and vitamins A, C, D, E. In the initial application of the TNI, based on dietary data from the 2011-2014 NHANES, the mean TNI score of U.S. adults was 75.4 out of 100, whereas the mean score when ignoring contributions from DS was only 69.0 (t-test; p<0.001). Therefore, the results from this analysis suggested that the TNI extends existing measures of diet quality by including nutrient intakes from all sources.</p> <p>However, another analysis designed to examine the construct and criterion validity of the TNI was completed using data from a nationally representative sample of U.S. adults who participated in the 2011-2014 NHANES, as well as exemplary menus reflective of high-quality diets. The results showed that the TNI yielded high scores on exemplary menus (84.8-93.3 out of 100), was moderately correlated with the Healthy Eating Index-2015 (r=0.48) and was able to differentiate between groups with known differences in nutrient intake for DS users vs. non-DS users, non-smokers vs. smokers, and those who were food-secure vs. food insecure (all <em>p</em><0.001). Moreover, correlations of TNI component scores with nutritional biomarkers for vitamins A, C, D, and E were significantly higher when compared with those obtained via the diet alone.</p> <p>Taken together, the research findings from the studies presented in this dissertation contribute evidence supporting the use of multiple modes of DS assessment to improve the comprehensiveness of DS exposure assessment, as well as the importance of capturing micronutrient exposures from all sources. These results also highlight the broad applicability of the TNI framework as a tool for comprehensively representing the totality of nutrient exposures, inclusive of DS. Our outcomes point towards a need for additional investigation into the random and systematic error that pervades DS assessment, the development of tools that can be used to describe micronutrient exposures more completely and how those relate to nutritional status, as well as further applications of the TNI. </p>

<p>3D scene reconstruction based on single-view images is an ill-posed problem since most 3D information has been lost during the projection process from the 3D world coordinates to the 2D pixel coordinates. To estimate the portion of an object from a single-view requires either the use of priori information such as the geometric shape of the object, or training based techniques that learn from existing portion sizes distribution. In this thesis, we present a single-view based technique for food portion size estimation.</p><p><br></p> <p>Dietary assessment, the process of determining what someone eats during the course of a day, provides valuable insights for mounting intervention programs for prevention of many chronic diseases such as cancer, diabetes and heart diseases. Measuring accurate dietary intake is considered to be an open research problem in the nutrition and health fields. We have developed a mobile dietary assessment system, the Technology Assisted Dietary Assessment^TM (TADA^TM) system to automatically determine the food types and energy consumed by a user using image analysis techniques.</p><p><br></p><p>In this thesis we focus on the use of a single image for food portion size estimation to reduce a user’s burden from having to take multiple images of their meal. We define portion size estimation as the process of determining how much food (or food energy/nutrient) is present in the food image. In addition to estimating food energy/nutrient, food portion estimation could also be estimating food volumes (in cm³) or weights (in grams), as they are directly related to food energy/nutrient. Food portion estimation is a challenging problem as food preparation and consumption process can pose large variations in food shapes and appearances.</p><p><br></p><p>As single-view based 3D reconstruction is in general an ill-posed problem, we investigate the use of geometric models such as the shape of a container that can help to partially recover 3D parameters of food items in the scene. We compare the performance of portion estimation technique based on 3D geometric models to techniques using depth maps. We have shown that more accurate estimation can be obtained by using geometric models for objects whose 3D shape are well defined. To further improve the food estimation accuracy we investigate the use of food portions co-occurrence patterns. The food portion co-occurrence patterns can be estimated from food image dataset we collected from dietary studies using the mobile Food Record^TM (mFR^TM) system we developed. Co-occurrence patterns is used as prior knowledge to refine portion estimation results. We have been shown that the portion estimation accuracy has been improved when incorporating the co-occurrence patterns as contextual information.</p><p><br></p><p>In addition to food portion estimation techniques that are based on geometric models, we also investigate the use deep learning approach. In the geometric model based approach, we have focused on estimation food volumes. However, food volumes are not the final results that directly show food energy/nutrient consumed. Therefore, instead of developing food portion estimation techniques that lead to an intermediate results (food volumes), we present a food portion estimation method to directly estimate food energy (kilocalories) from food images using Generative Adversarial Networks (GANs). We introduce the concept of an “energy distribution” for each food image. To train the GAN, we design a food image dataset based on ground truth food labels and segmentation masks for each food image as well as energy information associated with the food image. Our goal is to learn the mapping of the food image to the food energy. We then estimate food energy based on the estimated energy distribution image. Based on the estimated energy distribution image, we use a Convolutional Neural Networks (CNN) to estimate the numeric values of food energy presented in the eating scene.</p><p><br></p>

O gasto energético total (GET) pode ser usado como medida da ingestão energética (IE). Existe um constante sub-relato da IE obtida por métodos de avaliação do consumo alimentar, mas poucos estudos o investigaram em nações em desenvolvimento. Objetivos: a) comparar a validade das estimativas de IE de um questionário de freqüência alimentar, três recordatórios alimentares e um diário alimentar de três dias, segundo a água duplamente marcada; b) determinar a influência da prática de atividade física, do índice de massa corporal e de fatores psicossociais no sub-relato e; c) comparar as taxas de sub-relato entre agrupamentos de padrões alimentares. Métodos: Sessenta e cinco mulheres responderam aos métodos de inquérito supracitados, a partir dos quais foi estimada a IE. O GET foi medido pela água duplamente marcada. A prática de atividade física, índice de massa corporal, escolaridade, renda, idade, conhecimento nutricional, insatisfação corporal, restrição dietética, compulsão alimentar e o desejo de aceitação social foram correlacionados ao sub-relato. Os padrões alimentares foram obtidos pela análise de cluster. Resultados: O GET foi de 2.622 ± 490 kcal, enquanto que a IE, mensurada respectivamente pelo recordatório, diário e questionário, foi de 2.078 ± 430 kcal; 2.044 ± 479 kcal e 1.984 ± 832 kcal. A proporção de sub-relatores foi de 24,6% para o recordatório, 29,2% para o diário e 53,8% para o questionário (p < 0,005). Os sub-relatores apresentaram menores médias de renda e escolaridade e maiores valores de idade, insatisfação corporal e desejo de aceitação social. O sub-relato foi mais comum no padrão alimentar mais frugal. Conclusão: Os três métodos de avaliação do consumo alimentar apresentaram erros sistemáticos, embora o questionário de freqüência alimentar tenha tido o pior desempenho. O sub-relato foi influenciado por diversos fatores psicossociais e variou conforme o padrão alimentar relatado, o que pode comprometer a avaliação do consumo / Total energy expenditure (TEE) may be used as a measure of energy intake (EI). There is a constant underreporting of EI obtained by dietary assessment methods, but few studies have investigated it in developing nations. Objectives: a) to compare the validity of EI estimates obtained by a food-frequency questionnaire, three diet recalls and a three-day food record; b) to determine the influence of physical activity, body mass index and psychosocial factors on underreporting and; c) to compare underreporting rates between dietary pattern\'s clusters. Methods: Sixty-five women completed the dietary assessment methods, which were used to estimate EI. TEE was measured by doubly labeled water. Physical activity practice, body mass index, education, income, age, nutritional knowledge, body dissatisfaction, dietary restraint, binge eating and social desirability were correlated to underreporting. Dietary patterns were obtained by cluster analysis. Results: TEE was 2,622 ± 490 kcal, while EI, measured respectively by the diet recall, food record and food-frequency questionnaire, was 2,078 ± 430 kcal; 2,044 ± 479 kcal and 1,984 ± 832 kcal. Proportion of underreporters was 24.6% (recall), 29.2% (record) and 53.8% (questionnaire) (p < 0.005). Underreporters had smaller income and education and greater age, body dissatisfaction and social desirability. Underreporting was more common in the \'frugal foods\' pattern. Conclusions: The three dietary assessment methods presented systematic errors, although the foodfrequency questionnaire had the worst performance. Underreporting was influenced by psychosocial factors and varied according the reported dietary pattern, which may compromise dietary assessment

<p>The aim of the present thesis was to study elderly self-managing and disabled women’s dietary intake and meals in relation to age, household structure (single-living or cohabitant), disability and cooking ability. The women were aged 64-88 years and living at home, in the mid-eastern part of Sweden. The self-managing women were randomly selected. The disabled women – suffering from Parkinson's disease, rheumatoid arthritis or stroke – were selected from patient records. A total of 139 self-managing and 63 disabled women participated. Two dietary assessment methods were used: a repeated 24-h recall and a three-day estimated food diary, providing dietary intake for five non-consecutive days. The results indicate that elderly women still living in their homes seem to manage a sufficient dietary intake despite disability and high age. The reported energy intakes in all groups of women were low, which might be explained by an actual low intake and/or under-reporting. The portion sizes seemed to be smaller in the highest age group, leading to lower intakes of some nutrients. Thus also the nutrient density of the food should be given greater consideration. The meal pattern was shown to be regular and the distribution of main meals and snacks was found to be satisfactory. Meals and snacks that were defined as such by the women themselves thus seem to be more significant from an energy and nutritional perspective. Perceived cooking ability co-varied with energy and nutrient intake as well as with meal pattern.</p><p>Further, a qualitative dietary assessment method, FBCE, was analysed. It was concluded that it must be supplemented with a dietary assessment method providing energy intake figures to ensure a sufficient intake, especially when studying groups at risk for low energy intake. </p><p>Furthermore, the aim was to perform a dropout analysis. When studying older women and food, a low participation rate might be expected since the most active, the very ill as well as the disabled tend to decline participation, but also since food is a gender issue. Food could, especially for women, be a sensitive area of discussion, even though older women seem to choose "healthy foods" and eat "proper meals".</p>

The principal aim has been to investigate the contemporary role and importance of edible wild plants in the diets of women in different agro-ecological regions of Vietnam. Field studies were undertaken in four villages in the Mekong Delta (MD) and the Central Highlands (CH) in 1995-1999. Data collection included rapid appraisal techniques, botanical identification, dietary assessments (7-day food frequency recall), anthropometry, blood sampling (haemoglobin, serum ferritin, serum retinol and C-reactive protein) and analysis of vegetable samples (dry matter, selected minerals and vitamins, tannins and phytic acid). A food variety analysis was used to test the adequacy of diets. The uses of over 90 wild plant species were documented, many with multiple functions as foods, medicines and livestock feeds. In the dietary assessment 29 species were reported. Most women used a combination of home produced, commercial and wild vegetables. A high 72 and 43% respectively of the vegetables consumed in the MD and CH sites were gathered (rainy season). They contributed significantly to micronutrient intakes, especially vitamin A, calcium, vitamin C and folate. They added considerably to food diversity and women with the most diverse diets had relatively adequate nutrient intakes. The species used and importance of wild plants varied considerably with region, season and ethnicity. The main conclusions are that edible wild plants continue to make important contributions to the nutrient intakes of women. If this is neglected in diet assessments our understanding of the overall dietary adequacy may be misinterpreted and much valuable knowledge of traditional food diversification may be lost. A food variety analysis can be a useful tool in identifying groups with inadequate diets. The dual role of many species as dietary components and in preventive health care deserves greater attention.