Abstract: Various techniques are provided for performing real-time subject-motion-tracking during photon imaging by applying various data processing techniques on raw photon-events data generated by photon imaging scanners. In one aspect, a process of performing real-time subject-motion tracking during photon imaging begins by receiving multiple channels of raw singles-event data from a set of detector groups of the photon scanner while scanning a live subject. For each received channel of raw singles-event data, a singles-rate time series is generated based on a predetermined temporal resolution. Next, the set of singles-rate time series corresponding to the set of detector groups is combined to generate an overall singles-rate time series. Subsequently, the overall singles-rate time series is processed to extract in real-time one or more motion signals corresponding to one or more physiological motions of the live subject while the live subject is being scanned.

Abstract: A method, apparatus, and computer instructions stored on a computer-readable medium perform latent image feature extraction by performing the functions of receiving image data acquired during an imaging of a patient, wherein the image data includes motion by the patient during the imaging; segmenting the image data to include M image data segments corresponding to at least N motion phases having shorter durations than a duration of the motion by the patient during the imaging, wherein M is a positive integer greater than or equal to a positive integer N; producing, from the M image data segments, at least N latent feature vectors corresponding to the motion by the patient during the imaging; and performing a gated reconstruction of the N motion phases by reconstructing the image data based on the at least N latent feature vectors.

Abstract: A method, apparatus, and non-transitory computer-readable storage medium for image denoising whereby a deep image prior (DIP) neural network is trained to produce a denoised image by inputting the second medical image to the DIP neural network and combining a converging noise and an output of the DIP network during the training such that the converging noise combined with the output of the DIP network approximates the first medical image at the end of the training, wherein the output of the DIP network represents the denoised image.

Abstract: A method, system, and computer readable medium to perform nuclear medicine scatter correction estimation, sinogram estimation and image reconstruction from emission and attenuation correction data using deep convolutional neural networks. In one embodiment, a Deep Convolutional Neural network (DCNN) is used, although multiple neural networks can be used (e.g., for angle-specific processing). In one embodiment, a scatter sinogram is directly estimated using a DCNN from emission and attenuation correction data. In another embodiment a DCNN is used to estimate a scatter-corrected image and then the scatter sinogram is computed by a forward projection.

Abstract: To reduce the effect(s) caused by patient breathing and movement during PET data acquisition, an unsupervised non-rigid image registration framework using deep learning is used to produce motion vectors for motion correction. In one embodiment, a differentiable spatial transformer layer is used to warp the moving image to the fixed image and use a stacked structure for deformation field refinement. Estimated deformation fields can be incorporated into an iterative image reconstruction process to perform motion compensated PET image reconstruction. The described method and system, using simulation and clinical data, provide reduced error compared to at least one iterative image registration process.