乳腺肿瘤边界清晰度特征量化方法

摘要

The definition of breast tumor boundary often exists as an important feature for identification of malignant and benign breast tumor. The quantization of the definition of the breast tumor boundary benefits most of the methods that try to distinguish malignant and benign breast tumor. However, it is hard to quantify the definition of the breast tumor boundary by the naked eyes. We proposed an efficient algorithm to quantify the definition of the breast tumor boundary based on the consistency of gray level in the tumor region using ultrasound images and the result shows that our method performs as well as human. Thus this method can be seen as effective.%乳腺肿瘤边界清晰度特征是评判肿瘤良恶性的重要特征之一.对乳腺肿瘤边界清晰度特征的量化可以为基于该特征的乳腺肿瘤良恶性判别方法提供量化依据.而通过肉眼观察很难对乳腺肿瘤边界的清晰度进行有效的量化,因此,笔者提出一种基于区域灰度连续性的B超乳腺肿瘤边界清晰度量化方法.结果显示,利用该方法得到的乳腺肿瘤边界清晰度量化结果与通过肉眼观察的结果一致,说明这是一种有效的乳腺肿瘤边界清晰度量化方法.