抄録
The image reconstruction for diffuse optical tomography (DOT) employing a regularization method that minimizes entropy of the reconstructed optical property distribution was proposed to improve the contrast and noise robustness. We demonstrated the performance of the entropy regularization as compared to the l1-norm and l2-norm regularizations with numerical simulations and a phantom experiment. The entropy regularization provided the reconstructed images with improved contrast with artifacts caused by measurement noise suppressed. Our results indicate that entropy regularization has a potential to improve DOT imaging.