Abstract
The recently developed generalized linear least squares (GLLS) fast algorithm is useful in image-wide parameter estimation to generate parametric images with positron emission tomography (PET). However, parameter estimation using the GLLS may not converge for every pixel in noisy clinical PET data. In this study, a bound generalized linear least squares (B-GLLS) algorithm was proposed to guarantee that the estimation converges for noisy pixel curves and the parameters estimated are within the physiological and pathological ranges. The clinical results demonstrated that the B-GLLS algorithm can generate more accurate parametric images and is potentially useful in dynamic clinical PET studies.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
