INVESTIGADORES
DI PERSIA Leandro Ezequiel
congresos y reuniones científicas
Título:
Sparse Coding for apnea-hipopnea detection
Autor/es:
FEDERICO LESTUSSI; HUGO LEONARDO RUFINER; LEANDRO EZEQUIEL DI PERSIA; DIEGO HUMBERTO MILONE
Lugar:
Oro Verde
Reunión:
Congreso; XIV Reunión de Trabajo en Procesamiento de la Información y Control: RPIC 2011; 2011
Institución organizadora:
Universidad Nacional de Entre Ríos
Resumen:
The sleep apnea-hipopnea syn-drome is a common disorder which could bedetected by the decrease of the oxygen satu-ration. In this work we propose the use of asparse representation of the SaO2 signal to helpin the detection of this pathology. For trainand test we used the SHHS PolysomnographyDatabase which include 1000 polysomnograms.We show that the reconstruction of the SaO2signal from an optimal estimated dictionary haserrors in the order 10−13. The activation coeffi-cients obtained from this dictionary was used tonon-linearly regress the apnea-hipopnea marksby training a multilayered perceptron. Thenwe estimate the apnea-hipopnea marks from atest dataset with very good results, with a cor-relation coefficient of 81.74%.
