BIOLUMINESCENCE TOMOGRAPHY RECONSTRUCTION ALGORITHM BASED ON MULTI-TASK BAYESIAN COMPRESSED SENSING METHOD

摘要

Disclosed is a bioluminescence tomography reconstruction algorithm based on a multi-task Bayesian compressed sensing method, which belongs to the field of medical image processing. The method firstly uses the transmission rule of the high order approximation model of model light inside biological tissue, explores the inner correlation of a multi-spectrum based on a multi-task learning method, uses multispectral correlation as prior information and integrates same into a reconstruction algorithm so as to decrease the morbidity generated during BLT reconstruction, and finally realizes the three dimensional reconstruction of a fluorescent light source on this basis. Compared with other bioluminescence tomography reconstruction algorithms, the method of the present invention further has multispectral correlation integrated therein and decreases the morbidity generated during BLT reconstruction. The method can not only accurately relocate the fluorescent light source, but can also greatly improve the computational efficiency.