Automated Detection Method For Intra Cranial Lesions In CT Images

With the advent of multidetector computed tomography (CT) and fast scan times, CT has become the centerpiece for cranial imaging. It is the examination of choice for brain cancer, investigating stroke, intracranial hemorrhage, trauma and degenerative diseases. By observing and analyzing CT brain images, it may effective help doctor make accurate judgement to actual disease. However, the analysis on brain CT images, the course is dry as dust, need to deal with a large number of data every day, it is very apt to present the mistake. In this case, have proposed measuring, assisting and diagnosing, through some automatic methods automatically, cut apart, the characteristic is drawn, discern etc., get the area of pathological change, the doctor do one more careful inspection to area of these pathological changes finally, thus safe and more effective analysis patient's situation.Unlike the rich literature in brain segmentation from MRI data, research on segmenting brain from CT images is sparse. For that, this paper has done some researchs in computer-assisted automatic detection system of CT brain images, and proposed an algorithm based on SA-PSO and PCNN for segmenting CT brain images. Firstly, a 2D reference image is chosen to represent the intensity characteristics of the original 3D data set. Secondly, the ROI of reference image is determined as the space enclosed by the skull, then through this method based on SA-PSO and PCNN, the area of value of minimum gray level will be obtained by lighting a fire at the second time, combined ahead binary image with the third firing binary image, we can get another gray level image, this image may be some disease region. The next lighting binary image inclued the last one, so we can take the binary images of area of value of every gray level. Theses binary images are the regions of cerebrospinal fluid (CSF), gray matter (GM), and white matter (WM) and so on.