Method of images segmentation by dynamic spectral characteristics
Abstract
In considered method we use basic principles of biometrics and bioradiolocation for solution of the problem of images segmentation. Using dynamic spectral characteristics, obtained by means of wavelet spectrum we extract biometric indicator in form of signal of brightness pixels modification at the face skin part, specified by heartbeat. It is proposed to use quasi-matched wavelet filters for efficient selection of human heartbeat signal and it is shown the possibility of its frequency measurement practically at real-time mode. Obtained results can be used for many medical applications, security systems, object identification, etc.
Original Russian Text © A.A. Borodyn, 2016, published in Izv. Vyssh. Uchebn. Zaved., Radioelektron., 2016, Vol. 59, No. 1, pp. 55–60.
Abstract
In considered method we use basic principles of biometrics and bioradiolocation for solution of the problem of images segmentation. Using dynamic spectral characteristics, obtained by means of wavelet spectrum we extract biometric indicator in form of signal of brightness pixels modification at the face skin part, specified by heartbeat. It is proposed to use quasi-matched wavelet filters for efficient selection of human heartbeat signal and it is shown the possibility of its frequency measurement practically at real-time mode. Obtained results can be used for many medical applications, security systems, object identification, etc.
Original Russian Text © A.A. Borodyn, 2016, published in Izv. Vyssh. Uchebn. Zaved., Radioelektron., 2016, Vol. 59, No. 1, pp. 55–60.
Sign up for our newsletter
Get highlights and special offers twice a month
