灰度共生矩阵在尘肺阴影密集度判读中的应用

APPLICATION OF GLCM IN INTERPRETATION OF PNEUMOCONIOSIS SHADOWGRAPH INTENSITY

在实现尘肺病的自动分期判读过程中,针对X射线DR胸片受到尘肺病变多样性的影响导致无法直接通过测量阴影大小来准确获取阴影密集度的问题,提出基于灰度共生矩阵与BP神经网络相结合的尘肺密集度判读方法。通过Matlab对不同期别的尘肺样本进行仿真实验,结果表明由灰度共生矩阵产生的四个特征值能对不同期别尘肺胸片的纹理特征进行有效描述,并通过BP神经网络分类可实现对尘肺阴影密集度的有效判读。

In the process of realising automatic interpretation of pneumoconiosis in different stages, we propose the pneumoconiosis intensity interpretation method, which is based on the combination of GLCM （ gray level co-occurrence matrix） and BP neural network and aiming at the problem that the impact of pneumoconiosis lesions diversity on X-RAY DR chest radiograph leads to the impossibility of directly obtaining the shadowgraph intensity by measuring the size of shadow area. Simulation experiments of pneumoconiosis samples in different stages are carried out using Matlab, and the results show that the four eigenvalues generated by GLCM can effectively describe the texture fea- tures of the chest radiograph of pneumoconiosis in different stages. Moreover, it is able to realise the effective interpretation of shadowgraph intensity of pneumoconiosis through BP neural network classification.