基于边缘几何特征的乳腺X线图像中微小肿瘤检测方法

The Detection of Micro Tumor in Breast X-ray Image Based on Edge Geometric Features

乳腺癌是女性高发恶性疾病,乳腺X线图像是乳腺癌诊断的主要依据。目前乳腺X线图像中微小肿瘤检测方法的肿瘤边缘特征提取效果不佳,导致检测结果假阳性过高,影响病患的诊断与治疗,故研究基于边缘几何特征的乳腺X线图像中微小肿瘤检测方法。根据形态学原理,对乳腺X线图像进行增强处理。采用HOG特征提取方法获取乳腺微小肿瘤的边缘几何特征,并构建肿瘤边缘几何特征稀疏表示字典。使用边缘字典与差影技术,完成微小肿瘤检测。实验结果表明,与传统的检测方法相比,此方法可有效降低检测假阳性个数,能够提高微小肿瘤检测精度。因此,说明本文方法具有较高的可行性。

Breast cancer is a high incidence of malignant diseases in women and mammography is the main basis for diagnosis.At present,because of the tumor edge feature extraction of micro tumor detection methods in breast X-ray images is not effectively,the false positive results overhigh,which affects the diagnosis and treatment of patients.Therefore,the micro tumor detection method in breast X-ray images based on edge geometric features was studied.According to the morphology principle,the breast X-ray image was enhanced.HOG feature extraction method was used to obtain the edge geometric features of breast micro tumors,and constructed a sparse representation Dictionary of tumor edge geometric features.Using edge dictionary and subtraction technology to complete microtumor detection.The experimental results showed that compared with the traditional detection methods,this method could effectively reduce the number of false positives and improve the detection accuracy of micro tumors.Therefore,the method in this paper has high feasibility.