KAHIBARO
31.6. Projection Data
Table of Contents
Recording detected photons
In a gamma camera simulation, projection data represents how many photons are detected at each detector position for each view angle. In GATE you obtain this data by recording the energy deposits from incoming photons in the scintillation crystal and organizing them according to detector pixel or channel.
For a simple gamma camera, you usually already have geometry for the detector head, scintillation crystal, and collimator, and a digitizer chain that converts raw hits into singles. Projection data is then created by counting validated detector events in a 2D grid that corresponds to the detector surface.
A typical workflow is to attach an actor that records either energy deposition in a pixelized way or singles in a 2D image. In OpenGATE you can define an image-based actor positioned at the crystal or at the readout plane. The actor uses a grid of voxels that you align with the detector field of view. Each event that satisfies your energy window and threshold contributes to the corresponding detector bin, usually by adding 1 count or by accumulating detected energy.
For projection data, you are mainly interested in counts per detector element. The crucial configuration parameters are the image dimensions, that is the number of pixels in x and y, the physical size of the image covering the active detector area, and the quantity to record. You typically select counts to build a standard projection image, but sometimes energy-weighted images are also useful. Since projection data are 2D, the image depth is usually set to one voxel along the axis perpendicular to the detector plane.
During a SPECT acquisition, the detector head rotates around the source or phantom. At each projection angle, you either run a separate simulation with a fixed detector position and a dedicated output file, or you include the angle index as an additional index in your data. The most straightforward approach for beginners is to simulate each projection angle independently and produce one 2D image per angle, which you can later stack into a sinogram or a projection set in your analysis tools.
In practice you also combine projection recording with the digitizer configuration. The digitizer applies energy blurring and an energy window, so only photons that interact within the photopeak window are accepted as valid events. The projection actor then uses these accepted singles instead of raw hits. This matches how a real gamma camera forms its images and ensures that scatter and noise are treated consistently with the selected window.
Once the simulation finishes, the projection images are written to disk in an image format such as MHD or NIfTI, or in a simple array format. Each pixel value represents the number of detected photons at that detector position for the corresponding projection view. These projection datasets are the starting point for SPECT reconstruction and for analyzing count distributions, angular sampling, and system performance.
Views: 12