PET/CT图像质量主观评价与感知模型

Subjective assessment and perception model of PET/CT image quality

针对面向医学领域的主观评价数据库缺乏,导致医学图像质量评价(image quality assessment,IQA)算法性能难以分析的问题,基于双重刺激失真水平测试法,建立正电子发射断层显像/计算机断层扫描医学图像的主观评价数据库.对比13种国际通用IQA算法在数据库上的性能,分析不同退化方法对IQA算法的影响.结果表明,对新建立的图像评价数据库来说,特征相似性(feature similarity,FSIM)图像评价模型在相关性及稳定性方面明显优于其他IQA算法,包括目前医学领域主流的峰值信噪比评价指标.

Image quality assessment （ IQA） is highly dependent on subjective assessment. However, no subjective image quality assessment database is presently available for medical image, which poses a big challenge for the evaluation of medical IQA algorithms. In this paper, we propose the first positron emission tomography/computed tomography medical image subjective assessment database based on the double-stimulus impairment scale method. Performances of thirteen commonly used IQA algorithms are compared on the database. Moreover, effects of different image distortions on IQA algorithms are analyzed. Experimental results show that the feature similarity model outperforms other IQA methods, including peak signal to noise ratio, the most commonly used algorithm in the medical field.