基于激光图像处理技术的微小乳腺肿瘤精准诊断

Accurate Diagnosis of Minimal Breast Tumor Based on Laser Image Processing Technology

乳腺X线图像是乳腺肿瘤早期筛查诊断的主要方式,以提升微小乳腺肿瘤诊断精度为目的,为此研究了基于激光图像处理技术的微小乳腺肿瘤精准诊断方法。对初始乳腺X线图像进行图像预处理,提取乳腺组织;利用脉冲耦合神经网络对乳腺组织图像进行图像增强处理,提升图像亮度、突出纹理细节;对增强处理后的图像进行区域分割,通过多阈值方法为各微小乳腺肿瘤区域确定一个最优阈值,以此划分出乳腺肿瘤区域;利用归一化自相关系数计算乳腺肿瘤区域像素间相关性,确定乳腺肿瘤区域的纹理特征,根据微小乳腺肿瘤生长区域面积和边界周长确定乳腺肿瘤区域几何特征,根据纹理特征与几何特征诊断微小乳腺肿瘤性质。实验结果显示该方法对乳腺X线图像处理效果较好,诊断精度达到93.14%。

Breast X-ray image is the main way of early screening and diagnosis of breast cancer,in order to improve the diagnostic accuracy of micro breast tumor,this paper studies the accurate diagnosis method of Minimal breast tumor based on laser image processing technology.The initial mammography image is preprocessed to extract the breast tissue;the pulse coupled neural network is used to enhance the image brightness and highlight the texture details;the enhanced image is segmented,and an optimal threshold is determined for each small breast tumor region by multi threshold method,so as to divide the breast tumor region.The normalized autocorrelation coefficient is used to calculate the correlation between pixels in breast tumor region to determine the texture features of breast tumor region.The geometric characteristics of breast tumor region are determined according to the growth area and boundary perimeter of breast tumor,and the nature of breast tumor is diagnosed according to the texture and geometric features.The experimental results show that the method has a good effect on mammography image processing,and the diagnostic accuracy reaches 93.14%.