Stationary detector arrays paired with moving x-ray sources enable continuous 3D reconstruction, eliminating procedure delays from separate processing phases.
A classification system generates masked projections to isolate sub-object characteristics within compound objects.
Temporal voxel modeling reduces motion artifacts in computed tomography by applying dynamic basis functions to track kinetic parameters.
A multi-sequence scanning method merges overlapping helical CT paths into a single target area to optimize dose efficiency.
A reference image processing method separates signal regions to create a coil normalization map for MRI systems.
Weighted back projection corrects variable pitch spiral scanning data for accurate image reconstruction.
Segmenting PET raw data by movement status enables dynamic attenuation map application, resolving artifacts from patient motion during continuous scanning.
A filtering method preserves local averages in CT density measurements by distributing error to adjacent pixels.
Symmetry properties and SIMD techniques accelerate 3D PET image reconstruction, reducing hours-long processing to minutes while maintaining detection precision.
Replacing time coupling electronics with a coded aperture mask improves detection accuracy and reduces system costs in positron emission tomography.
A TOF-PET detector array uses a DMA rebinner card to map coincidence event data into projection space, eliminating discontinuous step-and-shoot scanning gaps.
Spectral detection provides compositional constraints that reduce sample preparation difficulty and accelerate image reconstruction.
A data processor segments body parts and calculates temporal gating functions for voxels to reconstruct three-dimensional images.
Segments vessel trees and maps time-series values onto voxels, enabling detailed aneurysm analysis beyond tubular structures.
Synthesizes subject projection data with phantom-derived noise to train machine learning models.
Gradient guidance in a GAN framework resolves clinical reliability issues by preventing fake details during ultra-high resolution CT reconstruction.
A convolutional neural network combines standard PET and MRI data to generate high-resolution images.
Automated image analysis detects body volumes and organ presence in CT scans, resolving the trade-off between manual verification accuracy and processing time.
A correction apparatus uses a linear absorption coefficient model to process projection images.
A trained non-linear mapping reconstructs low-quality ultrafast ultrasound images into high-quality outputs.
Iterative boundary estimation refines patient outlines to correct CT image reconstruction.
A garment pattern drafting system maps metric computational rules to two-dimensional image slices of a human form for precise panel dimensioning.
A parametric PET reconstruction method records bed position and time to calculate slice acquisition times for accurate image generation.
A neural network processes CT X-ray data to generate slice reconstruction images.
A helical computed tomography reconstruction method interpolates views into circular scan sinograms for backprojection operations.
Undersampled raw data reconstruction using iterative algorithms with offset slice geometry.