Mass Detection in Mammogram Based on Marker-pulse Coupled Neural Networks
-
摘要: 乳腺X线图像中的肿块检测是乳腺癌早期诊断的重要手段。该文提出了一种新的肿块检测方法。将脉冲耦合神经网络(Pulse Coupled Neural Networks, PCNN)与标记符相结合设计了标记PCNN图像分层方法，继而利用多同心层 (Multiple Concentric Layers, MCL)模型得到可疑区域。最后，借助肿块的形态学特征剔除假阳性区域得到最终的肿块。实验结果表明，该文方法在保证假阳性率(False Positive Rate, FPR)的同时，肿块真阳性率(True Positive Rate, TPR)达到92.08%。同时针对东方女性致密型乳腺案例中检测结果明显优于MCL方法和MCA方法。
-
关键词:
- 乳腺X线图像 /
- 乳腺癌早期诊断 /
- 肿块检测 /
- 标记脉冲耦合神经网络 /
- 多同心层模型
Abstract: Mass detection in mammogram plays an important role in early breast cancer diagnosis. A novel method of mass detection in mammogram is proposed. Combined with Pulse Coupled Neural Network (PCNN) model and marker-controlled watershed method, an image slicing method based on Marker-PCNN is presented. Then the suspicious regions are extracted though the Multiple Concentric Layers (MCL) analysis. Finally, the morphological features of mass are employed to eliminate the false positive areas. The experimentation results show that the detected method is excellent and the False Positive (FP) is low. The detection correction rate reached 92.08%. Compared with the original MCL method and Morphological Component Analysis (MCA) method, the proposed method has evident advantage, especially in diagnoses of dense breast cancer. -
计量
- 文章访问数: 2117
- HTML全文浏览量: 67
- PDF下载量: 734
- 被引次数: 0
下载:
