The fluctuations of the human pupil in presence of light stimulation have been long investigated in several clinical applications, both in natural and artificial conditions. The pupil dynamics offer useful information in order to make non-invasive diagnoses of neurological diseases. Typically the pupil is shot by a CCD camera, which is the core of the measurement apparatus, called pupillometer, and the resulting image is analysed. In this paper we present the application of a multiscale approach to edge detection to identify the morphological parameters of the pupil edge. First, we determine the degradation parameters of the measured image, which is assumed to be blurred by a Gaussian kernel and corrupted by an additive white noise; then we apply the edge detection procedure and the optimal fitting, showing the main results; a first dynamical analysis is also presented
Pupil edge detection and morphological identification from blurred noisy images / Iacoviello, Daniela; Lucchetti, M; Calcagnini, G; Censi, F.. - STAMPA. - 1:(2003), pp. 922-925. (Intervento presentato al convegno International Conference on the IEEE Engineering in Medicine and Biology Society tenutosi a Cancun nel 17-21 september 2003) [10.1109/IEMBS.2003.1280767].
Pupil edge detection and morphological identification from blurred noisy images
IACOVIELLO, Daniela;
2003
Abstract
The fluctuations of the human pupil in presence of light stimulation have been long investigated in several clinical applications, both in natural and artificial conditions. The pupil dynamics offer useful information in order to make non-invasive diagnoses of neurological diseases. Typically the pupil is shot by a CCD camera, which is the core of the measurement apparatus, called pupillometer, and the resulting image is analysed. In this paper we present the application of a multiscale approach to edge detection to identify the morphological parameters of the pupil edge. First, we determine the degradation parameters of the measured image, which is assumed to be blurred by a Gaussian kernel and corrupted by an additive white noise; then we apply the edge detection procedure and the optimal fitting, showing the main results; a first dynamical analysis is also presentedI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
