Federico Turco1, Irena Zubak2, and Johannes Slotboom1
1Institute of Diagnostic and Interventional Neuroradiology / SCAN, University Hospital Bern and Inselspital, University Bern, Bern, Switzerland, 2Neurosurgery, University Hospital Bern and Inselspital, University Bern, Bern, Switzerland
1Institute of Diagnostic and Interventional Neuroradiology / SCAN, University Hospital Bern and Inselspital, University Bern, Bern, Switzerland, 2Neurosurgery, University Hospital Bern and Inselspital, University Bern, Bern, Switzerland
Deep learning based metabolite quantification of in vivo MRSI data using CNN and ResNet were performed and compared to traditional NLLS-quantification. Accuracy measures were given, and both methods seem a viable alternative state of the art NLLS quantification.
Figure 3: Metabolite concentration mapping for Ace, Choline, and Creatine in rows 1, 2, and 3 respectively. Obtained by the three different methods, in (a) healthy brain spectra while (b) is a brain tumor.
Figure 1: Representation of both implemented neural networks, an CNN (a) and a ResNet (b). In both cases, the input and output are exactly the same, and all the convolutional layers have the same kernel size of 3.
