Chernoff faces: an orthodontic application

This retrospective study examined 21 children (12 males and nine females) with Class III malocclusions, who had been successfully treated by a non-extraction technique and for whom follow-up records were available at least 2 years after all treatment and retention had ceased. On the basis of their post-retention records, children were divided into two groups: those who remained stable (n = 10) and those whose treatment relapsed (n = 11). A graphical technique of data description, Chernoff faces, in which variables for each individual may be portrayed by appropriate scaling of selected facial features, was used to construct cartoon faces for each child. Seven variables, overjet, overbite, upper and lower incisor inclination, the numbers of anterior and posterior teeth in lingual occlusion and the angle ANB were chosen and diagrammatic faces drawn using data from both the start of treatment and post-retention stages. These were arranged according to relapse status and scrutinized to determine whether the groups possessed any common features which were related to the outcome of treatment. The faces showed a wide range of variation, especially at the pretreatment stage. Whilst the cartoons were easy to read, no typical features could be identified at the start of treatment which might indicate that therapy would later relapse.