MICCAI 2019

MICCAI 2019 DAILY Kerstin Kläser This combined network then minimises the pseudo CT error, but also the resulting PET residual. It does that by a combined loss. The first loss is the classical L2-loss and the second loss is the whole second network described above. It’s minimised as a whole. So, they minimise for the CT residual, but also for the resulting PET residual. Why is this important? Kerstin explains: “In order to do PET attenuation correction on a PET/MR, the gold standard is still to acquire an additional CT, which is not really healthy, and nobody really wants that. This is why CT synthesis in the first place is important. But for the sake of PET attenuation correction, it’s important that we minimise for the right thing. The right thing in this case is that we in the end generate a PET that is as close as possible to the real PET.” Current training on brain images is working well, and looking at next steps, Kerstin says they would like to try it on different and more difficult areas in the body. The lung in particular has huge problems with attenuation correction because there is no signal in the lung from the MR. She hopes that their work can achieve some improvement for this. 13 Little Ingrid is attending her first MICCAI with mom Aasa Feragen, who is supporting her student Rune Kok Nielsen, who has an oral on non-diffeomorphic image registration on Tuesday afternoon.