Nasal Cavity Detection in Facial Thermal Image for Non-contact Measurement of Breathing

In this research, we propose a method of improving on the accuracy of detecting nasal cavity location in far infrared images for non-contact measurement of human breathing. We found that although our previous method for far infrared imaging can detect regions that include nasal cavities well, several false alarms occur. In order to reduce false alarms, we propose to apply false alarm classification into our current method. Object detection method based on a boosted cascade of Haar-like feature classifiers are applied to find the candidates of the region including nasal cavities. In false alarm classification, binarize process is employed to segment facial area and background strictly. Based on the result of binarize process, false alarm on background can be classified from the results of detection. 5,100 FIR images are collected to train our nasal cavity detector; we evaluate the number of false alarms and detection failures. The results show that proposed method can reduce false alarm events.