ABSTRACT
Multiplexing is a way of increasing the sensitivity in a multi-slit-slat SPECT system by allowing the overlap of projections from neighboring apertures. The fundamental objective of multiplexing is to increase the signal-to-noise ratio for a given system resolution. Multiplexing may therefore lead to an improved tradeoff between resolution and sensitivity. Overlapped projections, however, introduce ambiguities in the data which can lead to non-uniqueness of solution for the inverse problem, deterioration in the quality of reconstruction and ultimately loss of resolution. Therefore, it is not straightforward to evaluate the advantage of the extra sensitivity gained by multiplexing, without first devising a method to overcome the image artifacts caused due to this overlapping of projection data. In this paper we investigate the effect of multiplexing on the reconstructed image quality and we determine whether reconstruction of multiplexed data could be improved by the addition of non-multiplexed data. For this purpose we have done simulations based on three digital phantoms. We compared the reconstructed images both qualitatively and quantitatively for different degrees of multiplexing and different fractions of non-multiplexed data. Our results indicate that the recovery coefficient (and therefore spatial resolution) can be maintained with a high degree of multiplexing leading to a significant increase in the SNR (up to 25%) due to a reduced noise level. This gain in the SNR corresponds to a 75% increase in counts or sensitivity which can be utilized to reduce acquisition time, patient dose or/and improve image quality.
