Search Results (939)

Background: Image-based dietary assessment provides a more intuitive approach for nutritional monitoring and health management. However, in multi-category bowl-based meals, food boundary adhesion, spatial stacking, and staple-food occlusion by upper-layer dishes still affect the accuracy of volume, weight, and nutritional composition prediction. Methods: This study proposes a nutrition prediction method for bowl-based foods by integrating semantic segmentation, multi-view three-dimensional reconstruction, and occlusion compensation. The improved DBP-FDSNet was used to extract food-category masks from top-view RGB images, while detail enhancement, boundary-assisted supervision, and spatial position encoding were incorporated to improve the segmentation quality of food boundaries and adhesion regions. The visible food surface inside the bowl was reconstructed using a bowl instance model and RGB-TSDF-based multi-view fusion, and the two-dimensional semantic results were mapped into the height-field parameter domain for category-level volume integration. For partially occluded, severely occluded, or completely invisible staple foods, a layered compensation strategy was introduced to reduce staple-food volume prediction errors and the erroneous assignment of upper-layer food volume. Food weight and whole-bowl Calories, Fat, Carbohydrate, and Protein were finally predicted using food density and a nutritional composition database. Results: DBP-FDSNet achieved a meanIntersectionoverUnion ($mIoU$) of 80.51% and a BoundaryF1 Score ($bF1$) of 85.73%. At the whole-bowl level, the MeanAbsolutePercentageError ($MAPE$) values for Calories, Fat, Carbohydrate, Protein, and total food mass were 13.23%, 18.51%, 14.18%, 13.35%, and 10.85%, respectively. Conclusions: The method improves the stability of category-level volume and nutritional composition prediction in complex bowl-based meal scenarios, providing a feasible solution for image-based dietary assessment and intelligent nutrition management. Full article

Background: Data on the nutritional adequacy of unsupplemented vegetarians in Northeast China are limited. Methods: We compared dietary intake, body composition, and serum biomarkers among vegans, lacto-ovo-vegetarians, and omnivores. This cross-sectional study included 356 adults (all diet ≥ 2 years, no supplements). Dietary intake was assessed using a validated semi-quantitative FFQ, body composition by BIA, and serum biomarkers. Kruskal–Wallis tests with Bonferroni correction were used. Results: Vegans had lower BMI (22.0 vs. 24.6 kg/m²), body fat (24.5% vs. 28.0%), and visceral fat (0.65 vs. 1.05 L) than omnivores (all p < 0.002). Vegans consumed more fiber (38.5 vs. 18.0 g/d) and vitamin C (180 vs. 85 mg/d), but less vitamin B₁₂ (0.3 vs. 4.2 μg/d), vitamin D (0.5 vs. 3.2 μg/d), calcium (520 vs. 720 mg/d), iodine (65 vs. 130 μg/d), and selenium (45 vs. 85 μg/d). Serum vitamin B₁₂, 25-(OH)D, ferritin, and selenium were significantly lower in vegans, while homocysteine was higher. The proportion of vegans with dietary intake below the recommendation reached 100% for vitamin B₁₂ and 97% for vitamin D, whereas omnivores showed excessive sodium (75%) and SFA (70%) intake. Conclusions: In this Northeast China cohort, unsupplemented vegetarian diets offered favorable macronutrient profiles and body composition but were associated with a high prevalence of dietary intakes below recommendations for vitamin B₁₂, vitamin D, iodine, selenium, zinc, and calcium. These findings underscore the need for targeted supplementation and food fortification strategies for individuals adhering to plant-based diets without supplement use in this region. Full article

Background: Ménière’s disease (MD) and tinnitus are common otological conditions that substantially impair quality of life. Although their pathophysiology remains incompletely understood, nutritional factors have been proposed to influence inner-ear microcirculation, water–electrolyte balance, oxidative stress and metabolic regulation. The objective of this scoping review was to comprehensively map the extent, nature and characteristics of existing research on nutritional strategies and dietary patterns applied in Ménière’s disease and tinnitus, and to clarify the mechanisms and clinical outcomes reported across studies. Methods: The review followed the Joanna Briggs Institute methodology and PRISMA-ScR guidelines. A comprehensive search of PubMed, Scopus, Web of Science, EBSCO, Cochrane Library and Google Scholar (2–12 March 2026) was conducted using the Population–Concept–Context model. Eligible studies included full-text human research (observational or interventional) and reviews published in English that examined dietary strategies in MD or tinnitus. Results: Of 273 records identified, 13 studies met inclusion criteria (6 on MD, 7 on tinnitus). Reported interventions included sodium reduction, adequate hydration, caffeine and alcohol modification, glycaemic stabilization, weight reduction, anti-inflammatory dietary patterns, and Mediterranean or DASH-style diets. Findings suggest potential symptom improvement in some patients. Conclusions: Available evidence indicates that nutritional interventions may serve as a valuable adjunct in the management of MD and tinnitus; however, their effectiveness has not been conclusively demonstrated. Well-designed, adequately powered randomised trials are still required to establish precise, evidence-based clinical guidelines. Full article

Background/Objectives: Accurate estimation of nutritional content from food images has important applications in dietary assessment and public health surveillance. While large language models (LLMs) have shown promise for this task, the effects of prompt design and model selection on estimation accuracy remain poorly characterized. Methods: We evaluated three Claude models (Haiku 4.5, Sonnet 4.6, Opus 4.6) for visual estimation of five mandatory nutritional components (energy, protein, fat, carbohydrate, and salt equivalent) across three datasets: NutriImage (691 Japanese meal photographs with dietitian-validated ground truth, after OCR-mask quality filtering), SNAPMe (1463 US meal photographs from a publicly available benchmark), and the Japan Branded Food Database (JBFD; 989–1000 packaged food product images). We systematically compared a default prompt and a visual estimation prompt explicitly instructing the model not to read any text or numbers visible in the image. Results: The visual estimation prompt substantially improved accuracy when paired with a sufficiently capable model (energy R²: 0.23 for Haiku to 0.60 for Sonnet, JBFD). Sonnet and Opus substantially outperformed Haiku across all datasets, while differences between Sonnet and Opus were small (MedAPE difference 1–3 percentage points). Packaged food images (JBFD) yielded higher R² than meal photographs. Salt equivalent showed consistently poor accuracy (MedAPE 34–64%). On SNAPMe, Sonnet achieved lower energy MAE (116.9 vs. 123.0 kcal, −4.9%) and lower MAE for protein (5.9 vs. 7.9 g, −25.7%) and fat (6.6 vs. 8.7 g, −24.5%) compared with a recent ChatGPT-5 study. Conclusions: Claude Sonnet offers the best cost-performance balance for LLM-based nutritional estimation. Prompt design substantially affects accuracy, but only when paired with a sufficiently capable model; model visual recognition capability appears to be a key determinant of performance. These findings highlight the inherent difficulty of this task and provide practical guidance for dietary assessment system development. Full article

Background/Objectives: Dietary assessment tools are essential for quantifying food and nutrient intake, characterising dietary patterns, and informing nutrition research. Food Frequency Questionnaires (FFQs) and Food Records (FRs) are widely implemented in observational studies, but completion rates vary, which may compromise data quality, introduce bias, and limit the interpretation of findings. This review is intended to synthesise evidence from observational studies on completion rates of these tools in populations with chronic conditions. Methods: A systematic search of Medline, PubMed, Scopus, and Google Scholar was completed. Eligible studies were observational studies using an FFQ or FR published from January 2015 to May 2025 in people with a chronic condition. Methodological quality was evaluated using the Hoy Risk of Bias tool for observational studies of prevalence. Subgroup meta-analyses estimated pooled mean completion proportions with 95% confidence intervals, and heterogeneity was assessed using the I² statistic. This study was registered with the International Prospective Register of Systematic Reviews and reported according to PRISMA guidelines. Results: Of the 8921 records screened, 88 studies (n = 84,579 participants) met inclusion criteria. The combined FFQ and food record mean pooled completion rate was 79.1% (95% Confidence Interval (CI): 74.37–83.43%). However, substantial heterogeneity was observed, indicating considerable variability across studies. Subgroup analyses highlighted important differences by tool type, format, age group, and disease category. FFQs demonstrated higher completion rates (80.6%) than FRs (74.3%). Electronic formats had higher completion rates than paper formats. Completion rates were higher in adults than in pediatric cohorts, and varied by chronic condition type, with kidney disease associated with the highest completion rates. Conclusions: These findings highlight the importance of considering tailored dietary data collection strategies, particularly for paediatric and medically complex populations, and provide direction for enhancing the feasibility of dietary assessment collection in future research. Full article

Background: Accurate dietary assessment is essential for estimating habitual and acute exposures. This study compared a 7-day food diary (7dFD) and repeated 24-h dietary recall (2×24hDR) for habitual and acute intake. Habitual fish and n-3 PUFA estimates were evaluated against whole-blood EPA and DHA, and acute intake of fish, rye bread, and coffee was explored using operational indicators. Methods: In a within-subject comparative design, 120 Danish adults aged 18–60 years completed both methods. Fish and n-3 fatty acid intakes were compared with whole-blood EPA and DHA. Usual fish intake from 2×24hDR was estimated using the Multiple Source Method (MSM) with Food Propensity Questionnaire (FPQ) data. Acute intake was assessed using single-day, consumption-day, single-meal, maximum-day, and meal-weighted estimates. Results: EPA, DHA, total fish, and fatty fish intakes estimated from the 7dFD and MSM-adjusted 2×24hDR correlated with biomarkers r = 0.23–0.46. Supplement inclusion improved EPA and DHA correlations (r = 0.26–0.59). The unadjusted 2×24hDR identified fewer fish consumers than the 7dFD (63% vs. 88%), whereas FPQ-based modelling identified 97%. Acute intake estimates varied by method and definition: the 2×24hDR produced higher short-term and upper- percentile estimates, while the 7dFD produced more stable distributions. Limitations include modest sample size, upper-percentile uncertainty, and non-equivalent supplement assessment. Conclusions: In this adult convenience sample, the 7dFD provided more stable habitual intake estimates, whereas the 2×24hDR produced higher short-term and upper-percentile estimates under the applied operational acute-intake definitions. These findings are context-dependent and should not be taken as evidence that either method is generally superior. Full article

Background/Objectives: Biological factors such as genetics contribute to nutrition-related outcomes, but nutritional epidemiological studies often lack consideration of genetics despite evidence of their functional impacts on health and cognition. Phenylalanine hydroxylase (PAH) genetic variation has been hypothesized to influence health and cognitive outcomes due to evidence of metabolic perturbations in L-phenylalanine to L-tyrosine hydroxylation, including plausible downstream effects on catecholamine neurotransmitters among not only individuals with phenylketonuria (PKU) [homozygotes for PAH mutations] but also PKU carriers [heterozygotes]. Related to these metabolic perturbations, diminished executive functioning has been observed in individuals with PKU, even when treated, but research is lacking exploring this outcome in PKU carriers. The present study aims to detail methods for stratifying populations based on genetic variation, for use in epidemiological precision nutrition research. It further provides an exploratory exemplar of such research through population stratification by PAH genetic variation (i.e., PKU carriers vs. non-carriers), while providing the first descriptive data on executive functioning skills using the validated Executive Skills Questionnaire—Revised (ESQ-R) tool with PAH-genetically stratified groups (PKU carriers and non-carriers). Methods: Participants were ≥18 years of age and PAH heterozygotes (PKU carriers) or non-carriers. Levels of executive functioning were self-reported anonymously online and included the validated Executive Skills Questionnaire—Revised (ESQ-R) tool. Data were analyzed using t-tests, chi-square tests, ANOVAs, and ANCOVAs. Results: Respondents (n = 99, n = 79 carriers and n = 20 non-carriers) consisted of males (22.2%) and females (77.8%), primarily of European ancestry. There were no significant differences between groups (carriers vs. non-carriers) for total scores (mean ± SD ESQ-R score carriers = 17.41 ± 14.01; non-carriers = 14.95 ± 10.00), but carriers scored significantly worse than non-carriers for the ESQ-R item “I have trouble making a plan” in the adjusted model. Conclusions: This study provides a methodological exemplar for exploring genetically stratified subpopulations in epidemiological precision nutrition research. Full article

Background/Objective: Rapid, validated dietary screening tools are lacking for individuals with spinal cord injury (SCI), where routine clinical check-ups do not allow sufficient time for extensive dietary assessments typically required to evaluate adherence to dietary recommendations. We developed a 15-item dietary screener (SCI NutriTool) and evaluated its accuracy in classifying non-adherence to a healthy food pyramid compared with a validated food frequency questionnaire (FFQ). Methods: The SCI NutriTool was developed through literature review and expert consensus. In a validation study, 51 adults with SCI (mean age 57.0 years; 76.5% men; 68.8% traumatic injury) completed the SCI NutriTool twice and a validated 97-item FFQ, which served as the reference method. Results: The SCI NutriTool demonstrated substantial variability in performance across food groups, reflecting its domain-specific screening properties. Sensitivity was high for fruits and vegetables (91.7%), protein-rich foods (90.5%), and sweetened/alcoholic beverages and snacks (82.4%), with relatively high positive predictive values (PPV: 73.7–90.5%), supporting the tool’s ability to identify individuals who are likely non-adherent and may benefit from further nutritional assessment or counselling. In contrast, for starchy foods and nuts, oils, and fatty spreads/sauces, sensitivity was low (20.0% and 50.0%), while specificity was modest. This indicates that the tool performs better in correctly identifying adherent individuals in these domains, which is reflected in higher negative predictive values (NPV: up to 94.1%). However, the low sensitivity suggests that individuals with non-adherence may be missed, limiting the tool’s usefulness as an early screening trigger for these food groups. Conclusions: The SCI NutriTool’s performance varies across food groups, demonstrating a stronger ability to identify non-adherence in protein-rich foods, fruit and vegetables, sweetened and alcoholic beverages, and snacks, but limited discriminatory capacity for others. In particular, it is not suitable for screening non-adherence to starchy foods and fats. Accordingly, it is best used as a triage tool to guide further dietary assessment and targeted nutritional interventions rather than as a standalone diagnostic instrument. Full article

Background and Objective: In recent years, there has been a dynamic development of artificial intelligence (AI), which has resulted in increased interest in its use in medical areas such as dietetics and psychodietetic. Current evidence remains insufficient to determine whether large language models (LLMs) are capable of generating diagnostically and therapeutically relevant responses in simulated psychodietetic scenarios. The aim of this study was to evaluate the performance LLMs—ChatGPT and Google Gemini—in simulated tasks related to diagnosis and psychodietetic intervention in anorexia nervosa (AN). Methods: Two complementary studies were conducted—the first, in which both models played the role of a patient suffering from anorexia and were tasked with answering questions asked in a psychodietetic interview, and the second, in which the chats analysed the case of a patient suffering from anorexia presented to them, and their task was to propose a correct diagnosis and therapy. Results: Both models have demonstrated the ability to engage with prompted scenarios and generate relevant, consistent responses. They were able to perform appropriate analyses, identify abnormal behaviors and formulate orderly diagnostic and therapeutic interpretations. At the same time, during the research, some differences were observed between the models in their interpretative approach and the level of development of the analyses. Conclusions: A key limitation of the models was a limited depth of empathic engagement and a reduced capacity for context-sensitive emotional responsiveness. Nevertheless, LLMs can potentially provide support in psychodietetic interventions if they are programmed correctly and are not abused in areas where they are not doing well enough. Full article

Background/Objectives: Patients receiving maintenance hemodialysis (HD) commonly exhibit chronic low-grade inflammation, nutritional disturbances, altered body composition, and metabolic imbalance. Gut dysbiosis may contribute to these abnormalities through the gut–kidney axis; however, the relationship between the gut microbiota composition and host phenotype in HD patients remains incompletely characterized. This study aimed to characterize the gut microbiota composition in maintenance HD patients and assess its cross-sectional associations with demographic, inflammatory, nutritional, dialysis-related, and bioimpedance-derived body composition parameters. Methods: This single-center cross-sectional study included 96 patients with end-stage kidney disease undergoing maintenance HD. The primary objective was to characterize the gut microbiota composition in maintenance HD patients. Secondary objectives were to assess cross-sectional associations with demographic factors (sex, age) and bioimpedance-derived body composition (specifically VAT). Clinical and laboratory data, inflammatory markers, nutritional indicators, malnutrition–inflammation score (MIS), dialysis-related variables, and bioimpedance-derived body composition parameters were collected. Stool samples were analyzed using full-length 16S rRNA sequencing. The gut microbiota composition was assessed using taxonomic profiling, alpha-diversity and beta-diversity analyses, subgroup comparisons, and exploratory distance-based analyses. Associations were interpreted within a descriptive and hypothesis-generating framework. Results: The gut microbiota composition showed marked inter-individual heterogeneity at the genus level, with dominant taxa including Blautia, Faecalibacterium, Streptococcus, Gemmiger, Ruminococcus, Escherichia-Shigella, and Enterococcus. Chao1 richness was higher in men than in women. Shannon entropy and Chao1 richness were positively associated with age and visceral adipose tissue (VAT), while Faith’s phylogenetic diversity increased with age. In contrast, the Gini index was negatively associated with age and VAT, indicating a more even microbial community structure in older individuals and in those with higher visceral adiposity. Beta-diversity analyses suggested modest differences in microbial community structure according to sex and selected body composition-related categories, particularly in sex-stratified analyses. Exploratory distance-based analysis showed a modest association between overall microbiota dissimilarity and host phenotype dissimilarity, although this finding was limited by reduced sample overlap. Conclusions: The gut microbiota composition in maintenance HD patients was highly heterogeneous and showed cross-sectional associations, mainly with sex, age, visceral adiposity, and broader host phenotype. These findings suggest that microbiota variation in HD reflects multidimensional demographic, inflammatory, nutritional, metabolic, and body composition-related factors rather than a single clinical determinant. Larger longitudinal studies integrating standardized dietary, medication, metabolic, and clinical outcome data are needed to determine the prognostic relevance of these microbiota patterns. Full article

Background/Objectives: Adolescence is a critical life stage characterized by rapid growth, increased nutrient requirements, and the establishment of long-term healthy behaviors. Growing evidence suggests that nutritional inadequacies may persist even when conventional indicators such as body mass index (BMI) appear normal, reflecting hidden malnutrition, a condition characterized by micronutrient inadequacy despite adequate energy intake. This issue may be particularly relevant in structurally constrained environments. This study aimed to compare dietary intake and nutrient adequacy between adolescents residing in residential care institutions (RCIs) and those in the general population in Taiwan. Methods: A total of 248 adolescents were included in the analysis. Institutional data were collected in 2018 and compared with nationally representative data from the Nutrition and Health Survey in Taiwan (NAHSIT 2010–2012). To improve comparability, 1:1 propensity score matching (PSM) was applied based on age, sex, and geographic region. Nutrient intakes were evaluated according to the Taiwan Dietary Reference Intakes (DRIs). Results: Adolescents in RCIs demonstrated significantly lower energy and protein adequacy than their counterparts in the general population. Among boys aged 13–15 years, the proportion meeting protein adequacy was substantially lower in RCIs than in the general population (34.0% vs. 84.0%). Similarly, among girls aged 13–15 years, energy adequacy was markedly lower in RCIs (25.0% vs. 63.9%). In addition, inadequate intake of multiple micronutrients, particularly B vitamins and essential minerals, was observed. Despite these differences, BMI remained largely comparable between groups, indicating a mismatch between anthropometric status and underlying nutritional quality. Conclusions: These findings suggest that hidden nutritional vulnerability may persist even within structured institutional environments designed to ensure stable food provision. The results highlight the limitations of relying solely on anthropometric indicators to assess nutritional status and underscore the need for targeted nutritional strategies to improve dietary quality and reduce health inequalities in residential care settings. Full article

Objectives: We aimed to develop a practical dietary quality score reflecting the Cantonese dietary pattern and evaluate its validity against established indices. Methods: The Cantonese Dietary Index (CDI, 岭南膳食指数) was constructed based on Cantonese dietary principles. Reliability was assessed using intraclass correlation coefficients (ICC) over 5–6 years in the GNHS. Validity was evaluated using Spearman correlations with dietary indices (aMed, DASH, and DBI) and by comparing associations with metabolic syndrome (MetS) across dietary indices using regression models. The CDI was developed from the Guangzhou Nutrition and Health Study cohort (GNHS) and validated in the Tianjin Chronic Low-grade Systemic Inflammation and Health (TCLSIH) cohort and the National Health and Nutrition Examination Survey (NHANES). Results: A total of 4025 (GNHS), 29,165 (TCLSIH), and 28,890 (NHANES) participants were included. Median CDI scores were 58.5, 51.0, and 49.0, respectively. The 5–6-year ICC was 0.33 (p < 0.001). The CDI was moderately correlated with dietary indices across the three studies (GNHS: from −0.55 [DBI-LBS] to 0.61 [DASH], TCLSIH: from −0.61 [DBI-DQD] to 0.71 [DASH], NHANES: from −0.33 [DBI-DQD] to 0.68 [DASH]). The odds ratios (95% CIs) of MetS for CDI, aMed, and DASH scores were 0.80 (0.74, 0.86), 0.91 (0.84,0.99), and 0.83 (0.77, 0.90) in GNHS, 0.95 (0.92, 0.98), 0.99 (0.96, 1.02), and 0.92 (0.89, 0.95) in TCLSIH, and 0.80 (0.77, 0.84), 0.80 (0.76, 0.84), and 0.72 (0.69, 0.76) in NHANES. Conclusions: The CDI demonstrated moderate validity and reliability in Chinese populations and was inversely associated with MetS. Full article

Background/objectives: The study aimed to identify the key methodological challenges and solutions related to recruitment and retention of rural-dwelling young adults into a randomized controlled trial that tests the feasibility of a digital healthy eating intervention (Veg4Me). Methods: Digital registration for a 12-week study was set up as a one-step process without researcher involvement. Participant registrations and recruitment rates were monitored daily using predetermined online preventative measures to identify fraudulent responses and to amend the digital registration process where necessary. Retention rates were monitored daily to identify any necessary amendments to the follow-up protocol. Results: During data collection, n = 279 fraudulent responses were identified from n = 536 total responses (52%). One month into recruitment, amendments were made to the registration process to reduce fraudulent responses. To address bot attacks, Qualtrics passwords and a two-factor authentication process were added to the Veg4Me landing page. Targeted recruitment strategies, such as unpaid social media posts, corresponded to peaks in recruitment. In the final recruitment month, a question was embedded within follow-up correspondence to encourage completion of the post-intervention survey. This resulted in an additional n = 8 (7%) participants completing the intervention. Conclusions: Empirical observations made in this study suggest that digital recruitment protocols without direct researcher involvement should consider multiple in-built strategies for identifying and preventing fraudulent responses. This includes a two-factor authentication process and minimizing the over-promotion of financial incentives in recruitment strategies. Recruitment strategies should consider the use of social media posts in local community groups, while the use of reminders and notifications could support retention. Full article

Background/Objectives: Understanding the microstructure of ingestive behavior (IB) is critically important to the development and success of interventions to change eating rates and produce more optimal food energy intake outcomes. Detailed measurement of IB microstructure is needed to guide development of real-time sensing approaches that can support such interventions. This article summarizes novel measurement and inference strategies around both digital video and inertial motion sensors in a structured laboratory protocol. Methods: Digital video footage was annotated for chews and bites and analyzed with generalized additive models to assess differences in IB for each of four meal courses varying by food texture. Results: Significant differences were revealed in IB microstructure in the form of nonlinear patterns of annotated video footage and initial sensing tests, indicating an optimal sensor location over the jaw’s condyle bone. Conclusions: Findings of an intensive longitudinal multicourse full meal protocol reflect important differences in nonlinear trends of eating behavior for diverse texture foods. These differences inform further development of technology-aided measurement strategies, provide an experimental protocol for fieldwide IB inquiry, and reveal expected fundamental differences in ingestion rates. Further inquiry into the underlying causes of nonlinearities for high UPF foods, along with sensor measurements, is warranted. Full article

Background/Objectives: Diet quality indices (DQIs) are commonly used to evaluate the alignment of dietary intake with national dietary guidelines. However, few methods have been developed to apply DQIs to image-based dietary assessment tools. This study aimed to develop a novel food-based Diet Quality Score (DQS) and examine its construct validity in Australian adults aged 18–65 years living with a higher weight (BMI 30–45 kg/m²), using the validated Healthy Eating Index for Australians 2013 (HEIFA-2013) for comparison. Methods: This cross-sectional study assessed dietary intake in 260 adults (mean age 48 ± 12 years; BMI 35.0 ± 4.2) collected over 4 days using the mobile Food Record (mFR™) and scored using the HEIFA-2013 and DQS (both scored out of 100). Descriptive statistics summarised participant characteristics with DQS and HEIFA-2013 scores. Construct validity was examined by assessing the associations between sociodemographic, health behaviour, anthropometric, and clinical blood marker variables and DQS and HEIFA-2013 scores. Multivariable binary logistic regression identified variables associated with being in the lowest diet quality tertile for both DQIs. Results: The mean scores were 47.4 ± 8.7 for the DQS and 52.0 ± 8.6 for HEIFA-2013. Both scores were positively associated with age (DQS: r = 0.216, p < 0.001; HEIFA-2013: r = 0.265, p < 0.001) and attention paid to the health aspects of their diet (DQS: p < 0.023; HEIFA-2013: p < 0.010). Increasing age (OR = 0.958, 95% CI: 0.932–0.985, p = 0.003) and moderate versus low physical activity (OR = 0.40, 95% CI: 0.21–0.80, p = 0.009) decreased odds of being in the lowest diet quality tertile for both DQI scores. Conclusions: The DQS demonstrated acceptable construct validity, providing a valid method for using image-based dietary assessments to evaluate diet quality in Australian populations. Full article