Positron emission tomography (PET) suffers from severe resolution limitations which reduce its quantitative accuracy. In this article, we present a super-resolution (SR) imaging technique for PET based on convolutional neural networks (CNNs). To facilitate the resolution recovery process, we incorporate high-resolution (HR) anatomical information based on magnetic resonance (MR) imaging. We introduce the spatial location information of the input image patches as additional CNN inputs to accommodate the spatially-variant nature of the blur kernels in PET. We compared the performance of shallow (3-layer) and very deep (20-layer) CNNs with various combinations of the following inputs: low-resolution (LR) PET, radial locations, axial locations, and HR MR. To validate the CNN architectures, we performed both realistic simulation studies using the BrainWeb digital phantom and clinical studies using neuroimaging datasets. For both simulation and clinical studies, the LR PET images were based on the Siemens HR+ scanner. Two different scenarios were examined in simulation: one where the target HR image is the ground-truth phantom image and another where the target HR image is based on the Siemens HRRT scanner — a high-resolution dedicated brain PET scanner. The latter scenario was also examined using clinical neuroimaging datasets. A number of factors affected relative performance of the different CNN designs examined, including network depth, target image quality, and the resemblance between the target and anatomical images. In general, however, all deep CNNs outperformed classical penalized deconvolution and partial volume correction techniques by large margins both qualitatively (e.g., edge and contrast recovery) and quantitatively (as indicated by three metrics: peak signal-to-noise-ratio, structural similarity index, and contrast-to-noise ratio).

中文翻译：

---

使用卷积神经网络的超分辨率 PET 成像

正电子发射断层扫描 (PET) 受到严重的分辨率限制，这降低了其定量准确性。在本文中，我们提出了一种基于卷积神经网络 (CNN) 的 PET 超分辨率 (SR) 成像技术。为了促进分辨率恢复过程，我们结合了基于磁共振 (MR) 成像的高分辨率 (HR) 解剖信息。我们将输入图像块的空间位置信息作为额外的 CNN 输入引入，以适应 PET 中模糊内核的空间变化特性。我们比较了浅（3 层）和非常深（20 层）CNN 与以下输入的各种组合的性能：低分辨率 (LR) PET、径向位置、轴向位置和 HR MR。为了验证 CNN 架构，我们使用 BrainWeb 数字模型进行了逼真的模拟研究，并使用神经影像数据集进行了临床研究。对于模拟和临床研究，LR PET 图像均基于西门子 HR+ 扫描仪。在模拟中检查了两种不同的场景：一种是目标 HR 图像是地面真实幻像图像，另一种是目标 HR 图像基于 Siemens HRRT 扫描仪——一种高分辨率的专用脑 PET 扫描仪。后一种情况也使用临床神经影像数据集进行了检查。许多因素影响所检查的不同 CNN 设计的相对性能，包括网络深度、目标图像质量以及目标和解剖图像之间的相似性。不过总的来说，