ULTRASOUND IMAGING PROJECT

Customer:
National Institutes of Health (National Cancer Institute)

Goal of Program:
We are evaluating a new algorithm for ultrasound B-scan image formation. The algorithm is based on a hybrid continuous-discrete state imaging model that represents the effects of abrupt changes in attenuation and scatterer density as the transducer pulse propagates across tissue boundaries and through cystic and calcified structures. We have tested our approach on phantom data and obtained substantial improvements in B-scan resolution and contrast. We expect our algorithm to result in improved imagery in a broad range of applications. We will focus our evaluation on transrectal prostate imaging with the goal of demonstrating significantly improved prostate zonal differentiation. The resulting imagery will enable physicians to better estimate prostate volume and identify the zones in which lesions are located or hypertrophy is occurring, capabilities important in the diagnosis of prostate-related medical problems.

Key Technology:
Our algorithm is based on a hybrid discrete-continuous state convolutional model that generalizes traditional convolutional models by taking into account changes in tissue reflection strength and attenuation. Our algorithmic structures are extremely efficient computationally, and will result in an overall algorithm that can be implemented in real-time hardware. Key Product: The ALPHATECH product resulting from this effort will be software that can be inserted into ultrasound imaging platforms through the use of commercially available digital signal processing boards.

Key Product:
The ALPHATECH product resulting from this effort will be software that can be inserted into ultrasound imaging platforms through the use of commercially available digital signal processing boards.