Accuracy in measurement of facial index on the basis of an android application

Overview

This study validates an Android application designed to measure the facial index, also known as the prosopic index, a metric used in orthodontic diagnosis to classify faces as broad, average, or narrow according to the Banister classification system. The facial index serves as a foundational measurement in comprehensive orthodontic treatment planning, with applications extending to anthropology and forensic medicine. The research evaluates whether photographic measurement via smartphone application can achieve accuracy comparable to traditional manual anthropometric methods using calipers. The investigation addresses the potential for digital tools to streamline clinical workflows by eliminating the need for specialized equipment and manual calculations while maintaining measurement reliability.

Methods and approach

The study employed a comparative design using a randomly selected sample of Indian subjects. Each participant's facial index was measured using two methods: the newly developed Android application, which utilizes smartphone photography, and traditional manual measurement with calipers. The statistical analysis employed an independent t-test to compare the facial index values obtained from the two methods. To assess the reliability of the application across different users, three separate examiners independently measured the same subjects using the app, with inter-examiner reliability evaluated through Cronbach's alpha coefficient. A Chi-square test was applied to compare the distribution of prosopic index classifications between the two measurement methods, examining whether subjects were consistently categorized into the same facial phenotype classifications regardless of measurement technique.

Results

The comparison revealed no statistically significant difference between facial index values derived from the Android application and those obtained through manual measurement. Inter-examiner reliability analysis yielded a Cronbach's alpha value of 0.998, indicating exceptionally high agreement among the three different investigators using the application. The facial phenotype distribution showed that the majority of subjects in the sample were classified as hyperleptoprosopic, with this pattern consistent across both the application-based and manual measurement methods. The concordance in classification outcomes between methods suggests that the digital approach successfully replicates the discriminatory capacity of traditional anthropometric techniques for categorizing facial types according to the Banister system.

Implications

The validated Android application offers a practical alternative to conventional anthropometric measurement in clinical orthodontics, eliminating the requirement for calipers and manual calculation while maintaining measurement accuracy and reliability. The high inter-examiner reliability indicates that the application can produce consistent results across different operators, reducing measurement variability that may occur with manual techniques. This technology represents a potential advancement in clinical efficiency for orthodontic practices, particularly in settings where rapid assessment is beneficial or where standardization of measurement protocols across multiple practitioners is desired. Beyond orthodontics, the application may have utility in anthropological research and forensic identification contexts where facial classification is required. The successful validation in an Indian population sample suggests applicability across diverse populations, though further validation in other ethnic groups would strengthen evidence for universal clinical adoption.

Disclosure

• Research title: Accuracy in measurement of facial index on the basis of an android application
• Authors: Rohan Pulgaonkar, Agrima Thakur
• Publication date: 2026-01-13
• DOI: https://doi.org/10.25259/anams-2022-3-1-(565)
• OpenAlex record: View
• Disclosure: This post was generated by artificial intelligence. The original authors did not write or review this post.