尘肺病数字胸片计算机辅助检测方法研究-生物医学工程专业论文.docxVIP

幅图像平均分成训练集与测试集，训练集与测试集各包含 300 幅正常胸片 以及 125 幅尘肺胸片。通过对训练集的训练，得到六个子区域的分类器参 数模型及其权重因子。然后用训练集得到的参数来评估测试集的分类性能。 本文不仅测试了全部特征数据集的性能，还测试了经过特征筛选后数据集 的性能，并且用两种不同的胸片分类器进行分类。结果显示特征选择大大 提高了计算机自动处理效率，节省了大量时间。全部特征数据集的分类敏 感性为 92.0％，特异性为 87.7％，准确性为 88.9％，经过 ROC（Received Operating Characteristic）曲线分析，得到 AUC（Area Under the ROC Curve） 值为 0.978。筛选特征数据集分类敏感性为 91.2％，特异性为 86.3％，准确 性为 87.8％，AUC 值为 0.958。 结果表明这个系统可以显著地将正常胸片 与尘肺胸片区分开来，满足临床尘肺病体检筛查的要求，为医生，更是为 病人提供了一种好的解决方案。 关键词：尘肺病 数字胸片 计算机辅助检测 主动形变模型 纹理分析 支持 向量机 COMPUTER-AIDED DETECTION FOR PNEUMOCONIOSIS ON DIGITAL CHEST RADIOGRAPHY ABSTRACT Pneumoconiosis is a serious occupational disease on the world wide with high incidence. The computer aided detection for pneumoconiosis is meaningful to high volume screening, which helps doctors optimize the diagnosis workflow, reduce workload and improve detection sensitivity. This paper presents an automatic computer aided detection scheme on digital chest radiographs to detect pneumoconiosis. Firstly, the lung fields are segmented from a digital chest X-ray image by using the active shape model method. Then, the lung fields are subdivided into six non-overlapping regions, according to Chinese diagnosis criteria of pneumoconiosis. The multi-scale difference filter bank is applied to the chest image to enhance the details of the small opacities, and the texture features are calculated from each region of the original and the processed images, respectively. After extracting the most relevant ones from the feature sets, support vector machine classifiers are utilized to separate the samples into the normal and the abnormal sets. Finally, the final classification is performed by the chest based report-out and the classification probability values of six regions. Experiments are conducted on randomly selected images from our chest database. Both the training and the testing sets have 300 normal and 125 pneumoconiosis cases. In the training phase, training models and weighting factors for each region are d