Table 7.

Classification of scatter correction methods.

Category	Methods Description	Benefit	Limitations	Examples
Hardware-based techniques	If coarse septa or beam stoppers are used. lines of response intercepted by the septa can be used to determine the scatter component	No noise increase	Unused	[442,443]
Multiple-energy window techniques	The energy spectrum is estimated by using windows below and above the photopeak window	Highly used Simple	Noise	[444,445]
Convolution and deconvolution-based techniques	In this case, the standard energy acquisition window is used. Data collected in it helps to estimate the distribution of scatter	Good image contrast Good accuracy	Not commonly used	[446], [447], [448]
Direct calculation techniques	Extract information from emission data, or a combination of emission and transmission data for estimating scatter distribution. Monte Carlo technique and ToF information can achieve great progress	The most popular High accuracy	Computational cost	[449], [450], [451], [452]
Iterative reconstruction-based scatter-correction techniques	Scatter distribution is obtained and used during image reconstruction	Parallel processing High contrast Low noise	Computational cost	[368,[453], [454], [455]