The Restoration and Correction of Photoacoustic Tomography for MRI Information

Photoacoustic tomographic (PAT) is a new kind of molecular imaging that can precisely map endogenous contrast agents and the physiological metabolism of tissues. PAT imaging's ultrasonic detecting method has made pinpointing particular lesions a long-standing difficulty. This is complicated and prone to error since the PAT picture has weak contrast between hard and soft tissues. To get over this problem, this study first makes use of the rich structural information obtained from another state-of-the-art imaging method, magnetic resonance imaging (MRI), to support organ segmentation and make up for the loss of light fluence brought on by PAT. We find that the suggested CNN approach is more successful in removing deviations from 2D PAT photos retrieved from sparse data compared to the standard UNet after contrasting the quality of restored pictures using both methods. An iterative optimization-based restoration framework incorporating two types of regularizations is developed for use in picture restoration. We test the suggested approach on phantoms as well as in vivo mice imaging, and the findings demonstrate a significant increase in picture quality and resolution.