Personal Recognition using Head-Top Image for Health-Monitoring System in the Home

Automatic health-monitoring systems for the smart house are being developed for the elderly. An automatic health-monitoring system needs a way of personal recognition when two or more aged persons live together. We propose a personal recognition method based on the space spectrum of the head-top image. We examined 33 head-top images from eleven subjects and achieved a personal recognition rate of 86.4 percent. When one subject with thinning hair was excluded, the personal recognition rate was 90.0 percent in 30 head-top images from ten subjects.