Nose detection in far infrared image for non-contact measurement of breathing

Nose Detection and Breathing Monitoring in Thermal Images

“…Nose hole detection had a miss rate of 8.6%, a hit rate of 12.8%, and a false alarm rate of 81.6%. These figures were similar to the results obtained by Hanawa et al, [12].…”

“…Thus, it is inappropriate to apply algorithms for visual images to thermal images [2]. A few studies have begun developing algorithms for nose detection in thermal images [11,12]. Al-Khalidi and his colleagues used the Prewitt operator to detect the boundaries of the face in the first thermal image, and then located the eyes and nose in the identified face [11].…”

“…Hanawa and his colleagues directly located the nose in thermal images by using Haarlike features and the Adaboost algorithm developed by Viola and Jones [12][13]. They used thermal images of 9 subjects recorded for 10 seconds each at 2700 frames.…”

Nasal cavity detection in facial thermal image for non-contact measurement of breathing

“…Face-detectors using their method have become popular because of their speed and robustness. In this research, we replace the gray-scale training images with FIR images for nose detection and replace gray-scale input image with an FIR image for detection [1]. GentleBoost is used here to learn classifiers in our method.…”

“…Far Infra-Red (FIR) imaging is receiving attention as an attractive way of realizing this function. We have been studying a machine learning algorithm to detect the nose in thermal images of the user's face [1]. We found that while our previous method can detect regions that include nasal cavities well, the false alarm rate is not insignificant.…”