γ刀定位中SPECT图像的自动处理方法研究

Automatic Processing of SPECT Image During Orientation of γ-Knife

目前γ刀手术通常都是根据MRI进行定位 ,但对于原发性癫痫 ,MRI无法显示 .将SPECT功能图像与MR解剖图像融合是目前最有前途的一种手术治疗原发性癫痫的定位方法 .但因SPECT图像与MRI不匹配 ,所以需要对SPECT图像进行处理 .根据SPECT图像中癫痫病灶的灰度分形特征 ,利用分形理论 ,改进了盒维数计算方法 (RDBC) ,对图像进行分形维数分割 .提出了自动选取型值点的算法 ,将分割后的病灶进行边缘拟和 .通过处理SPECT图像 ,使之与MRI融合 ,从而对癫痫病灶精确定位 ,指导伽玛刀手术 .MR和SPECT费用相对较低 ,因此该技术的开发 ,必将进一步推动伽玛刀治疗的普及 .

Orientation is usually achieved according to MRI during operations using γ-Knife at present,but MRI can not show natural epilepsy.Fusion of SPECT image with MRI is one of the ways to achieve orientation during any epilepsy operation.Most of SPECT images do not match with MRI,and so SPECT images must be processed.Accoroding to the fractal characterisics of epilepsy in SPECT images,RDBC is improved to divide the dimension and an algorithm is proposed for automatically choosing value points of the image.The processed SPECT image is fused with MRI to guide the γ-Knife during operation.