Facial dysmorphologies have shown potential as diagnostic factors for genetic and mental disorders. While they are easily detectable in certain conditions, they are more subtle in disorders like autism and psychotic disorders. To fully utilize facial biomarkers for diagnosis, affordable tools for accurate 3D facial shape acquisition and automated algorithms for detecting and computing facial biomarkers are needed. This would enable the development of low-cost and non-invasive clinical tools, reducing diagnosis time and costs while benefiting patients, families, and the healthcare system. However, current technology for 3D facial shape capture is complex and expensive, and the manual computation of facial biomarkers requires expert anatomists. This hinders the implementation of facial biomarkers in clinical practice. To address this, the proposal suggests developing a low-cost approach that uses depth sensors on smartphones to automatically compute facial biomarkers from 3D scans of patients' faces. The solution involves creating a functional application for clinicians to enter patient data, obtain 3D facial scans using smartphone depth cameras, upload the models to the cloud along with clinical metadata, and deploy trained deep learning models for accurate 3D facial landmarking through edge computing.