Skip to main content
EJNMMI Physics
Fig. 3 | EJNMMI Physics

Fig. 3

From: 90Y SPECT scatter estimation and voxel dosimetry in radioembolization using a unified deep learning framework

Fig. 3

Architecture of our residual learning network (DblurDoseNet) for SPECT-based absorbed dose-rate map generation. Each blue box corresponds to a multi-depth (the depth is listed on top of each box) feature map. The depths of kernels in the three 3D-convolution operations are 5, 5, 3, respectively. The input size was 512 × 512 × 11 with 11 adjacent slices to capture the dose contributions from 90Y beta particles. The spatial dimensions are maintained at 512 × 512 in all layers

Back to article page