Segmented attenuation correction using artificial neural networks in positron tomography.
Phys Med Biol
; 41(10): 2189-206, 1996 Oct.
Article
en En
| MEDLINE
| ID: mdl-8912390
ABSTRACT
The measured attenuation correction technique is widely used in cardiac positron tomographic studies. However, the success of this technique is limited because of insufficient counting statistics achievable in practical transmission scan times, and of the scattered radiation in transmission measurement which leads to an underestimation of the attenuation coefficients. In this work, a segmented attenuation correction technique has been developed that uses artificial neural networks. The technique has been validated in phantoms and verified in human studies. The results indicate that attenuation coefficients measured in the segmented transmission image are accurate and reproducible. Activity concentrations measured in the reconstructed emission image can also be recovered accurately using this new technique. The accuracy of the technique is subject independent and insensitive to scatter contamination in the transmission data. This technique has the potential of reducing the transmission scan time, and satisfactory results are obtained if the transmission data contain about 400,000 true counts per plane. It can predict accurately the value of any attenuation coefficient in the range from air to water in a transmission image with or without scatter correction.
Buscar en Google
Bases de datos:
MEDLINE
Asunto principal:
Tomografía Computarizada de Emisión
/
Redes Neurales de la Computación
/
Fantasmas de Imagen
/
Circulación Coronaria
/
Corazón
/
Modelos Estructurales
Tipo de estudio:
Prognostic_studies
Límite:
Humans
Idioma:
En
Revista:
Phys Med Biol
Año:
1996
Tipo del documento:
Article
País de afiliación:
Canadá