Department of Thyroid and Breast Surgery, East Branch of Quanzhou First Hospital, Fujian 362000, China.
Quanzhou First Hospital Affiliated to Fujian Medical University, Quanzhou, Fujian 362000, China.
Math Biosci Eng. 2023 Jan 4;20(3):4896-4911. doi: 10.3934/mbe.2023227.
Breast cancer occurs in the epithelial tissue of the gland, so the accuracy of gland segmentation is crucial to the physician's diagnosis. An innovative technique for breast mammography image gland segmentation is put forth in this paper. In the first step, the algorithm designed the gland segmentation evaluation function. Then a new mutation strategy is established, and the adaptive controlled variables are used to balance the ability of improved differential evolution (IDE) in terms of investigation and convergence. To evaluate its performance, The proposed method is validated on a number of benchmark breast images, including four types of glands from the Quanzhou First Hospital, Fujian, China. Furthermore, the proposed algorithm is been systematically compared to five state-of-the-art algorithms. From the average MSSIM and boxplot, the evidence suggests that the mutation strategy may be effective in searching the topography of the segmented gland problem. The experiment results demonstrated that the proposed method has the best gland segmentation results compared to other algorithms.
乳腺癌发生在腺体的上皮组织中，因此腺体分割的准确性对医生的诊断至关重要。本文提出了一种用于乳腺 X 线摄影图像腺体分割的创新技术。在第一步中，算法设计了腺体分割评估函数。然后建立了一种新的突变策略，并使用自适应控制变量来平衡改进的差分进化（IDE）在调查和收敛方面的能力。为了评估其性能，将所提出的方法在许多基准乳腺图像上进行了验证，包括来自中国福建泉州第一医院的四种类型的腺体。此外，还对所提出的算法与五种最先进的算法进行了系统比较。从平均 MSSIM 和箱线图来看，证据表明，突变策略在搜索分割腺体问题的地形方面可能是有效的。实验结果表明，与其他算法相比，所提出的方法具有最佳的腺体分割结果。
