Abstract
Recent advancements in mobile augmented reality (AR) and machine learning (ML) have enabled new approaches to visualizing genetic traits through facial phenotyping. This study presents an overview of Flutter framework for developing a cross-platform mobile AR application designed to display ML-predicted genetic facial features in real time. The proposed architecture integrates components for facial tracking via ARKit and ARCore, model inference using TensorFlow Lite, and dynamic heatmap visualization within a Flutter-based user interface. The study outlines the core components of the application, including frontend UX layers, local inference modules, and optional cloud-based backend infrastructure. Emphasis is placed on ensuring consistent user experience (UX) across platforms, privacy-aware data handling, and interpretable output for non-technical users such as medical professionals and students. Practical implementation cases are discussed, including preprocessing, mobile inference optimization, and face-aligned visual overlays for interactive genetic trait feedback. Additionally, challenges such as performance on low-power devices, user digital literacy, and model explainability are addressed. The integration of AR and ML into a unified mobile system opens new possibilities for clinical screening, education, and the broader adoption of assisted diagnostics.