Study On Techniques Of Computer Aided Detection Of Pulmonary Disease

With the exacerbation of environmental pollution, lung cancer and other pulmonary diseases continue to rank as the leading reason of death, and significantly debase the quality of lives. The early detection and treatment is the most effective way to prevent incurability of pulmonary diseases. CT scan, which is one of the most common imaging methods, the number of CT images is increasing in an exploding mode. This fact has lead to burdensome work loaded for physicians and negative influence of diagnostic performance. However this problem can be solved with the help of computer, which is also called after Computer Aided Detection, due to the maturation of Digital Image Processing, Pattern Recognition and Machine Learning theory and technologies. To avoid the mistakes in human brain 3D reconstruction by the physicians, a method of 3D Computer Aided Detection is necessary.The 3D Computer Aided Detection of pulmonary diseases is studied in this thesis. The main aspects of research include:1. The pre-processing method, which contains the filtering of volume data and the interpolation of volume data, are adopted in this thesis. In the research on volume data filtering, anisotropic diffusion algorithms are discussed deeply. A gradient direction-based anisotropic diffusion is proposed. In this algorithm, the value of K is calculated according to the histogram of gradient first. And then gradient directions are classified based on the features of itself. At last, diffusion coefficients are computed independently. In the research on the interpolation of volume data, an linear interpolation algorithm is applied and the pixels of the interpolated image derived by the algorithm are more practical for the applications.2.In the process of three-dimensional detection of suspected pulmonary diseases, Level set algorithm and Mean Shift algorithm with 3D region growth, which developed rapidly in recent years, are introduced to change the current situation that the traditional pulmonary diseases detection method are overused. In the detection, lung tissue must be segmented. A Live Wire algorithm with Canny operator is proposed to overcome the lose of edges detected by Gaussian operator. It is more suitable for medical image segmentation. And the results of this algorithm is well.3. Three algorithms for volume rendering, such as Raycasting, Splatting and Shear-warp, are carefully studied in this thesis to visualize the result of detection. After the experiment of six datasets, a conclusion is made by considering the facts of memory usage, CPU usage, the rendering time and image quality that Shear Warp is fit for daily diagnosis because of the lower system resources cost and the good image quality, while Splatting is fit for accurate use such as surgery plan and so on.4. 3D Computer Aided Detection system needs to provide the doctor fine experience of interactive.3D interactive methods with mouse are researched in the paper because the interactive exclusive used equipment is expensive. The 3D measurement is realized, using the algorithms of acupuncture pick and 3D rotation. The combine visualization of suspected diseases and normal tissues provide more clear evidence to the doctor. Another kind of virtual reference tool is diagnostic biopsy. A virtual plane is created first in the space, make it cut through dataset which will be processed. After computed, data information cut through by plane will be mapped to the plane. That is displaying the interested part.The experimental results of filtering, segmentations, volume rendering, and 3D interactive are presented in this thesis. These experimental results indicate that the 3D Computer Aided Detection of suspected pulmonary diseases, 3D measurement can be well done by proposed algorithms in the thesis. Also human issue can be displayed respectively according to the idea of physicians.