Lung Ct Image Recognition Based On Classification

With the development of medical image diagnostic technology, a large number of medical images are produced. Conventional chest radiograms have been used to screen lung and computed tomography is widely used to help detect lung images, because through this, it is possible to diagnosis correctly. However, this process is exhausting for there are so many images that should be interpreted. In addition, the error of reading could not be avoided. Therefore, computerized automated recognition for medical image could be used to help clinicians diagnose, treat, monitor changes, and plan and execute procedures more safely and effectively.In this thesis, an approach based on classification technology lung CT images recognition is studied and implemented, which can identify the suspect focus after the automatic segmentation of CT images, and greatly helps doctors improve the diagnostic quality and efficiency. Firstly, CT-image character and the CT-image-obtaining problem are solved by the study of the structure of DICOM file, which is used to store lung CT image. Then, the lungs image is preprocessed using different filter approach and the image intensification technology. Through threshold segmentation way and the region segmentation way's research, an improve the approach is proposed, which carries on the comparison with the original segmentation approach, then extracts image the region of interest, is using the region growing approach to carry on processing to the region, outlines the lung parenchymal of the lung CT image. Finally BP neural network is carried on classification through the extracting lung CT image characteristic, and the classified test is done.The experiments indicated to the lungs CT image show that the approach proposed in this thesis has the high retrieval efficiency and the validity to the lungs CT image recognition, the ease of operation.