颅内异常声压干扰下脑瘫病变区域分割仿真

Intracranial Cerebral Palsy Lesion Area Segmentation Under Abnormal Sound Pressure Disturbance Simulation

研究脑瘫病变医学图像边缘准确分割问题。脑瘫病变医学图像受到颅内声压的干扰,在采集过程中对病变区域特征属性形成干扰,造成病变区域特征边缘模糊化。传统算法在随机声压干扰下的脑瘫病变医学图像特征边缘模糊,无法形成有效的约束,造成病变区域分割精度下降。为提高精度,提出利用GCA演化模型的脑瘫病变区域分割方法。采集脑瘫病变区域图像,计算图像中不同像素的关联性,为克服干扰,对图像像素进行碰撞运算,实现脑瘫图像边缘增强。计算SUSAN检测算子,获取GAC演化模型,对获取的结果进行抗干扰处理,实现脑瘫病变区域分割。实验结果表明,利用改进算法进行脑瘫病变区域分割,能够避免颅内异常声压的干扰,为临床诊断提供依据。

Research on cerebral palsy lesions medical image recognition accuracy optimization problem. By the interference of intracranial pressure, cerebral palsy lesions in medical images in the lesion area feature attributes in the process of acquisition, cause the lesion area feature edge blur. Traditional algorithms in this random sound pressure under the disturbance of cerebral palsy medical image features of edge blur, can＇t form effective constraint, causing the deterioration of lesion region segmentation accuracy. Therefore, based on evolution of GCA cerebral palsy lesion region segmentation method of the model. Image acquisition cerebral palsy lesion area, calculating the correlation of different pixel in image, in order to overcome interference, the image pixel collision operation, realize the cerebral palsy image edge enhancement. Calculation, SUSAN detection operator, obtain GAC evolution model, the obtained results are anti-interference processing, realize the cerebral palsy lesion area division. Experimental results show that the improved algorithm for cerebral palsy pathological changes of regional segmentation, to avoid the abnormal intracranial pressure, to provide basis for clinical diagnosis.