Sound synchronization and motion compensated reconstruction for speech Cine MRI.
Pierre-Andr Vuissoz 1,2 , Freddy Odille 1,2 , Yves Laprie 3,4 , Emmanuel Vincent 3,5 , and Jacques Felblinger 6,7
Imagerie Adaptative Diagnostique et
Interventionnelle, Universit de Lorraine, Nancy,
INSERM, Nancy, France,
Universit de Lorraine, Nancy, France,
CNRS, Nancy, France,
INRIA, Nancy, France,
Hospital Nancy, Nancy, France,
1433, INSERM, Nancy, France
To construct an articulatory model of the vocal tract 10
sentences were pronounced and recorded at 3T using
balanced-SSFP multiphase sequences with a resolution of
256x256. An optical microphone was used to record the
speech during an 80 second scan. A noise reduction
algorithm is applied before aligning each acoustic
sentence onto a median template sentence. A cine loop of
128 images for each sentence is reconstructed using a
motion compensated sliding window algorithm (80 ms)
producing a reconstructed frame rate above 100 Hz. Each
cine loop enables the delineation of the vocal tract
with sufficient spatial and temporal resolution.
This abstract and the presentation materials are available to members only;
a login is required.