Analysis of facial expressions using thermal imaging

摘要

The following paper presents the development of a system for analyzing facial expressions using thermal imaging. The human subject is monitored during the test using a thermal camera. The recorded frames from a time unit are converted into images that can associate in a time period with the other biomedical signals acquired. The thermal images are decomposed using wavelet type of algorithms, using different levels of decomposing and associated functions. Since wavelet transformations act like filters, the resulting images will highlight certain areas of the face and implicitly the temperature changes. Using the correlation functions, we compared the images from the database built on patterns with the images obtained in real time.