Automated quality assessment of body MR images in the context of the German National Cohort MRI study

图像质量评价是影像学研究的重要组成部分。在大型流行病学成像研究中，包括德国国家队列（NAKO） MRI研究，由于收集的数据量很大，无法由人类读者进行这种质量评估。拟议研究的目的是在NAKO MR研究的背景下开发和实施一种身体MR图像自动质量评估的方法。为此，将根据NAKO MR协议实验获取人体MR图像，并在受控的实验条件下改变图像质量。产生的不同质量的图像将由放射科医生评定。基于这些评级，将识别图像特征，使用机器学习方法可靠地预测图像质量。所实施的方法将在NAKO MR研究中使用。

Assessment of image quality is an important part of imaging studies. In large epidemiological imaging studies, including the German National Cohort (NAKO) MRI study, this quality assessment cannot be performed by human readers due to the large amount of collected data.The aim of the proposed study is to develop and implement a method for automated quality assessment of body MR images in the context of the NAKO MR study.To this end, body MR images will be experimentally acquired according to the NAKO MR protocol and image quality will be varied under controlled experimental conditions. Resulting images of different quality will be rated by radiologists. Based on these ratings, image features will be identified that reliably predict image quality using machine learning methods.The implemented method will be used in the context of the NAKO MR study.