A Multimodal Cardiovascular-Safety-Aware System for Personalized Diet and Exercise
Personalized diet and exercise applications com-monly rely on demographic and anthropometric information such as age, gender, height, weight, body mass index, activity level, fitness goals, and food preferences. Although some systems inte-grate wearable data, heart rate is frequently used for monitoring rather than as a transparent safety constraint in recommendation generation; blood-pressure information, body-pose analysis, and non-diagnostic facial-skin observations are also rarely integrated into one explainable decision pipeline. This paper proposes a multimodal, cardiovascular-safety-aware mobile health framework that generates adaptive diet and exercise recommendations using anthropometric characteristics, optional body and facial images, heart-rate monitoring, and repeated blood-pressure readings. The system contains six modules: user profiling and health screening; body-pose analysis using MediaPipe Pose; optional facial-skin analysis for acne-like lesion severity estimation; cardiovascular risk and safety classification; constraint-based diet planning; and adaptive exercise recommendation. The facial-image component is intentionally non-diagnostic: it identifies only visible acne-like skin features and does not infer digestive disorders, nutritional deficiencies, hormonal disorders, or internal-organ conditions from facial location. Heart rate and blood pressure are used as primary safety signals to determine exercise intensity, progres-sion, alerts, and relevant dietary constraints such as lower-sodium alternatives where repeated elevated blood-pressure readings are observed. The proposed system will be evaluated through an eight-week pilot study involving adult participants. Outcomes will include model performance, recommendation relevance, usability, adherence, heart-rate and blood-pressure safety events, and selected fitness indicators. The framework is intended to support health and fitness planning and does not replace diagnosis, treatment, or professional medical advice.
Authors
- Dr.Abuzar Ansari Shruti Devdas Salvi
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-10-21
- DOI
- https://doi.org/10.5281/zenodo.23511984
- Primary Topic
- Mobile Health and mHealth Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00