Uso de técnicas de inferência estatística no estudo de padrões térmicos faciais

Abstract

This work consists in the study of facial thermal patterns of several volunteers in order to verify the repetition of such patterns to each individual and if they differ from one individual to the other. These data can be utilized individually or as an auxiliary method in several applications like: individual recognition of masked people, identification of health or emotional variations, etc. To study the viability of temperature marks use as an individual and inherent part of one, such as the digital print, temperature samples were collected from a volunteers' group. The collected temperature marks were selected so as to map parts of easy location on the face and they were not based on the shape or any physiological parameter of the face. Subsequently these samples were earlier processed so that methods of statistical inference would be used. The best result of this work is the one from the male volunteers' group with no sickness accounts, which samples confirm with 78% that a person's thermal pattern repeats itself. The statistical methods utilized in this work were not able to divide the groups so as to identify each group as having a different pattern from the others.