融合感知损失的深度学习在常规MR图像转换的研究

目的:研究在完全无监督的条件下深度神经网络实现常规磁共振图像间相互转换的可行性。方法:在循环生成式对抗网络(CycleGAN)中引入感知损失,使网络利用对抗损失学习图像结构信息的同时,结合循环一致性损失和感知损失生成高质量的磁共振图像,并将生成图像与CycleGAN模型以及有监督的CycleGAN模型(S_CycleGAN)生成的图像进行定量比较。结果:引入感知损失后的网络生成的图像定量评估值均高于CycleGAN模型生成的图像,生成的T1加权图像(T1WI)的定量评估值也均高于S_CycleGAN模型生成的T1WI,生成的T2加权图像(T2WI)与S_CycleGAN模型生成的T2WI的定量评估值相似。结论:在CycleGAN中引入感知损失,可以在完全无监督的条件下生成高质量的磁共振图像,进而实现高质量的常规磁共振图像的相互转换。     

Attitude Toward the Elderly,Fear of Death,and Work Preference of Baccalaureate Nursing Students

Aging increases the proximity to death. It may be that fear of death has a negative impact on attitude toward the elderly. The previous research is inconclusive. A positive attitude toward the elderly in past studies was not found to be related to preference for working with this age group. In the present study of senior baccalaureate nursing students, no relationship was found between attitude toward the elderly and fear of death. However, a significant relationship was found between positive attitude toward the elderly and preference for working with this age group. In addition, subjects who had experience living with or caring for an elderly person were found to have a high preference for working with the elderly. An intriguing finding is that subjects with a high preference for working with older clients with more likely to describe themselves as religious or devoutly religous.