Positron emmision tomography image reconstruction

RILRecon is used for image reconstruction from PET coincidence data. Image reconstructions are performed using the maximum-likelihood (ML) or penalized maximum-likelihood (PML) method. The orthogonal distance-based ray-tracer (OD-RT) with a Gaussian kernel is used as the geometrical projector. GPU acceleration is implemented based on NVIDIA CUDA C programming. The PML method is designed for mitigating limited-angle artifacts in the image reconstruction for a PET system that does not have full angular coverage of the imaging plane. A whole-body PET image of the same object will be needed for regularization. Users have the option to incorporate an image-based resolution model into the regularization term for better image quality. [Source]


Simulation environment and double Q-learning code for automated radiation source detection

Current radiation source survey strategies either require human efforts or are not efficient and flexible enough to adjust survey paths based on recent measurements. Reinforcement learning, which studies the problem of how agents ought to take the optimized action so that a goal can be achieved efficiently, provides an alternative data-driven solution to conduct radiation detection tasks with no human intervention. This code provides the simulated radiation environment and a double Q-learning algorithm for automated source searching. [Source]