Wavelet-based regularisation for background correction within the sub-millimeter 3D MLEM list-mode reconstruction process for high resolution small animal PET data

The measure data obtained from the acquiring PET system tend to be very noisy, since the limitation of the current instrumentation, detector efficiency, as well as random and scatter contamination are strong sources of noise. Therefore, the reconstruction method should include an estimation of the statistical nature of the noise. In small animal images, this problem is severe, corrupting areas of interest within small organs. In this paper, a Wavelet-based regularisation within the list-mode MLEM reconstruction process method is proposed, with the aim to regularise counts. Image reconstruction using a ¹⁸F small animal NEMA phantom and a ¹⁸F rat bones were performed. Feasibility of the method is demonstrated with and without wavelet-based regularisation correction. For each reconstruction case, investigation on the effects of image quality will be addressed. Results show that the proposed method gives an important counts regularisation as well as an improved background correction and contrast recovery.