Reliability of Comprehensive Facial Soft Tissue Landmark Detection and Analysis Using Frontal View Photographs.

Area of Science:

• Facial analysis and anthropometry
• Orthodontics and maxillofacial surgery
• Forensic science and identification

Background:

• Facial soft tissue analysis is crucial for orthodontics, orthognathic surgeries, and forensic anthropology.
• Assessing facial symmetry reliability in frontal photographs requires standardized methods.
• Evaluating landmark detection consistency across observers and time is essential for accurate analysis.

Purpose of the Study:

• To determine a standardized midline for reliable symmetry analysis in frontal facial photographs.
• To assess the consistency of soft tissue landmark detection and analysis between different observers and over time.
• To identify the most reliable midline for symmetry assessments in facial analysis.

Main Methods:

• Utilized 50 standardized frontal photographs of Iranian patients (aged 9-40) undergoing orthodontic treatment.
• Annotated 37 facial landmarks to perform 73 distinct facial analyses.
• Employed a Python-based automated framework and assessed inter- and intra-observer reliability using Intraclass Correlation Coefficient (ICC) and Mean Absolute Deviation (MAD).

Main Results:

• The perpendicular of the interpupillary line through the glabella emerged as the most reliable midline for symmetry analysis.
• Facial analyses demonstrated generally good reliability, with notable exceptions.
• Lower reliability was observed for landmarks like Gonion (Go) and Malar Eminence (ME), impacting specific width and symmetry analyses.

Conclusions:

• Established a comprehensive collection of soft tissue landmarks, facial analyses, and their reliability metrics.
• Identified the optimal facial midline for robust symmetry analysis.
• Highlighted that 3D-defined indices may have reduced reliability when applied to 2D images, impacting derived analyses.