Image Reconstruction Of Panel PET Based On Depth Of Interaction Information

Position Emission Tomography (PET) is one of the most important imaging tools in the field of medical imaging. From the level of molecular in living body, PET gives functional information about the body and plays an important role in early diagnosis and therapy of cancer.Traditional PET is ring and can not popularize widely because of the high cost. Compared to ring-PET, the panel-PET is cheaper, higher flexibility and have advantages in the specific organs of the human body and small animals. Therefore, the structure in this thesis is panel-PET.But annihilation photon incident at oblique angles on the detector leads to the parallax errors,it severely degrade the spatial resolution and the uniform of the reconstructed images. Traditionally, parallax errors have been addressed by making the detector ring much larger than the field of view (FOV). But it increases the cost of the system and reduces the sensitivity. The depth of interaction (DOI) researched in this thesis can mitigate parallax errors on the basis of low cost and high sensitivity to improve the spatial resolution and uniform of the reconstructed images.In this thesis, two structures of panel-PET are designed and proved that the reconstruction based on DOI information is better than traditional reconstruction in the image quality obviously. In addition, the reconstructed results and spatial resolution is different because of the different structure. For the structure of two panels, the crystal thickness is 5mm. If the DOI information cannot be précised to 5mm, the spatial resolution of reconstructed results cannot be improved almost. If the DOI information can reach 5mm, the spatial resolution of reconstructed results improve greatly. For the structure of four panels, if the DOI information can reached 4mm, the spatial resolution improved greatly, otherwise changed slowly. It tell us that if we use the DOI information, the DOI information must be reached to some degree, otherwise we cannot improve the spatial resolution of reconstructed results through the DOI information obviously. The reconstructed results also point out that only the DOI information is smaller than some value, the image quality improved greatly, otherwise the image quality cannot be improved along with the DOI information become bigger.