ART-TV Algorithm for Diffuse Correlation Tomography Blood Flow Imaging

Abstract

Near-infrared diffuse correlation imaging (DCT) is an important method of tissue blood flow imaging for the prognosis and diagnosis of various diseases. A new solution of DCT that is based on the Nth-order linear (NL) algorithm, termed as NL-DCT, was proposed in our previous study to overcome the limitations of tissue geometry and heterogeneity. The NL-DCT converts the image reconstruction into linear equations, and this solution is an ill-posed problem in mathematics. To improve the accuracy and robustness of the DCT image reconstruction, a combination of algebra reconstruction technique (ART) and total variation (TV), namely ART-TV, is proposed in this study. After each ART iteration, the TV model is used as an a priori constraint to reduce noise. The validations from computer simulation and phantom experiments with different anomalies demonstrate that the proposed ART-TV algorithm is efficient in DCT blood flow image reconstruction.

Keywords

• Diffuse correlation tomography
• blood flow index
• image reconstruction
• Nth-order linear approach
• algebra reconstruction technique
• total variation