由于心脏舒张、收缩以及血液的流动，CT心脏图像易出现弱边界、伪影，传统分水岭算法易产生过分割，为此，提出一种双边滤波和标记分水岭相融合的CT心脏图像分割方法（BF-WS）。采用双边滤波算法对心脏图像进行平滑滤波，并采用形态学对图像进行重构，以消除图像噪声，保留边缘信息，采用改进Otsu算法提取CT心脏图像的内、外标记，并采用分水岭分割算法实现CT心脏图像分割，采用临床CT心脏图像在Matlab平台进行性能测试。结果表明，BF-WS提高了CT心脏图像分割准确率，与专家分割结果十分接近，较好地解决了传统分水岭算法过分割难题，可以为临床医学诊断提供重要依据。%Because the cardiac, systolic and diastolic blood flow, CT cardiac images appear weak boundary, artifacts, the traditional watershed algorithm is easy to produce segmentation, therefore, the paper proposes a segmentation method of CT cardiac image based on Bilateral Filtering and the Watershed(BF-WS). The bilateral filter algorithm is used to filter for cardiac image, and the images are reconstructed by morphology to eliminate the image noise and preserve the edge information, and then the labeling of CT cardiac images is extracted by improved Otsu algorithm, and CT cardiac image is segmented by watershed segmentation algorithm, cardiac image of clinical CT images is used to test the performance in the Matlab platform. The results show that the proposed has improved cardiac CT image segmentation accuracy, and is very close to the segmentation results of experts, can solve the over-segmentation problem of the traditional watershed algorithm, and can provide an important basis for clinical diagnosis.
展开▼
