Abstract:

Lung nodes are frequently observed in cases of cancer. Nowadays, nodes can be viewed using technologies such as tomography or MRI. However, X-ray imaging is a low-cost method as well as its widespread use. In this context, the use of X-ray images in frequent intervals of monitoring the developments of the nodes is beneficial in many ways. This study suggested a three-stage innovative approach that can automatically track the development of nodes from lung röntgen images. In the first stage, the images taken at different times were carried out in a row to assess the development of a patient's nodules. The homographic matrix was obtained using the methods of self-quality extract and matching to be able to make upper-range sliding and then the image collision process was carried out. In the second phase, the nodes on the collapse images, according to the previously known location information, identified the corresponding nodes in both images. It is estimated that the nodes that do not match in the first picture have disappeared, only the nodes in the second picture have just emerged. In the final stage, the nodes are considered as a closed field consisting of a piksel set on the image and the field information that the closed fields form after matching is calculated. In this way, the conditions of growth or reduction are multiplied. The patient data set provided by the Department of Radiology of the Capital University has been used to test the developed approach. The test results were confirmed with the support of a specialist radiologist. The nodular development tracking system presented in this study has achieved hopeful results.

Keywords: