Détection automatique de micro-emboles cérébraux grâce à un nouveau détecteur de variance conditionnelle

The detection of precursor signs of cerebrovascular accidents is a problem of public health since cerebrovascular accidents correspond to the second leading cause of death worldwide. The transcranial Doppler system is used to detect micro-emboli signals. To date, the commercial devices used by clinician did not permit to detect all micro-emboli. New detectors are expected to reduce the false negative rate. We propose a novel detector based on the conditional variance. This detector outperforms the conventionnal detector by a factor 10. Consequently, it will allow a more efficient detection of precursor signs of cerebrovascular accidents and a reducing of the risk for patients.