Compression of a Monte-Carlo based system matrix for iterative reconstruction of PET images

Modern human PET-scanners often have high scatter fractions due to the lack of septa. This work treats the scatter in a straight forward manner by calculating the system matrix elements directly with Monte-Carlo (MC) methods. A parametric compression method was used to scale down memory consumption. The resulting images (reconstructed with either the compressed or with the uncompressed MC-matrix) were compared.