Automatic Detection Technique And Clinical Applications Of The Chest X-ray Of The Digital Images (dr) In Small Pulmonary Nodules Is Discussed

Imaging diagnosis is possessed of very important in diagnosis of the breast disease. Being regarded as one of the general clinical examination, the traditional chest radiography is still ubiquitous in clinical practice. Yet, its interpretation is notoriously difficult because of the limit of the film/screen system.Recently, digital radiography, especially the application of the DR (direct digital radiography) system leads the traditional radiography into digital field. To compare with traditional radiography image, direct digital radiography image have higher quantity and include further more information. After image manipulation, the digital information can be applied in further more clinical diagnosis. Along with the rapid develop of the technique of computer and image manipulation, the application of the technique of the CAD (computer-aided diagnosis) in the medical image gradually take to clinic.This paper reviews the development of the technique of CAD on chest image and the application of main technique of the digital image manipulation. Based on this, this paper brings forward the DR-based technique to make application research about small nodules automatic detection on digital radiography; the main is about detection of lung edge to obtain the ROI (region of interest), and then detect the small lung nodules automatically in the ROI so that it can find the pneumonic cancer early. On the other hand, we do a analysis of ROC (receiver operating characteristic) curve about the clinical diagnostic result of the radiologists to the cases that include 100 digital chest images with nodules and normal respectively in the clinical application experiment. The results of clinical diagnosis without and with CAD system is that the average area under the ROC curve (Az) value for residents and radiologists respectively are from 0.857 to 0.951 and from 0.892 to 0.959.This shows that radiologists can improve their rate of diagnosis by applying CAD system in the clinic. And the residents without experience may obtain so much from this. The better results of experiment inspirit us to make further researches. And then, we expect to develop the application software that can filter the nodules automatically and rapidly and accurately in chest DR images take from a mass screening.