DEVELOPMENT AND EVALUATION OF RECONSTRUCTION METHODS FOR AN IN-HOUSE DESIGNED CONE-BEAM MICRO-CT IMAGING SYSTEM

Purpose: Cone-beam micro computed tomography CT (micro-CT) can offer comprehensive 3D information and high-resolution images. This technology can be used with other imaging modalities in the preclinical application of molecular imaging research. Image reconstruction algorithm affects final image spatial resolution, which is the primary topic of this study. We use two types of reconstruction methods, which are analytical (tent-FDK) and statistical iterative (convex algorithm) algorithms, to reconstruct the micro-CT images for evaluation. Materials and Methods: In this study, projection data of the Defrise phantom and HA (Hydroxy-Apatite) phantom were obtained using an in-house designed micro-CT imaging system and images were reconstructed with the tent-FDK and convex algorithms. We develop a new way to calculate the system matrix of our micro-CT. Human tooth sample and mouse bone sample data were reconstructed with the analytic and iterative algorithms. Results: The Defrise phantom results show coronal view of the images reconstructed by the tent-FDK and convex algorithms. From the profile of the results, image reconstructed by the convex algorithm has higher pixel value in the high-density layer. Comparison of the results of human tooth sample and mouse bone sample reconstructed by the two kinds of algorithms, the image reconstructed by the convex algorithm has fewer artifacts and more correct pixel value representing for the attenuation coefficients. Conclusion: Tent-FDK algorithm is a kind of useful analytic method to reconstruct cone-beam CT data. We also develop the convex iterative method to reconstruct images for low dose X-ray cone-beam CT, which performs better than the analytical one in general.