Extraction Of Head Bone Tissue Based On CTA Images

With the rapid development of our country, the desire for healthy living of nationals is increasing. Cardiovascular and cerebrovascular disease is a serious threat to human health, in particular to the elderly over the age of 50. Nearly 3 million people die from cardiovascular and cerebrovascular disease each year, accounting for 51% of China’s total annual death. Therefore, the early diagnosis and accurate diagnosis of cardiovascular and cerebrovascular disease becomes an important topic in modern medicine. Computed tomography angiography (CTA), in medicine also known as non-invasive vascular imaging techniques, now has become an important diagnostic method for cardiovascular and cerebrovascular disease, and especially plays an irreplaceable role in the interventional treatment.This thesis aims to more accurately and more rapidly separating and extracting of the head bone tissue, assist doctors in doing a more accurate diagnosis. Firstly, according to the physiological characteristics of the human body and the characteristics of head medical images, a medical image pre-processing algorithm and a feature region skeletal statistics head hierarchical algorithm is proposed. The proposed medical image pre-processing algorithm can effectively remove the unrelated part in medical images, and retain useful parts of the body in the same time. The proposed head hierarchical algorithm can divide the complex head data set into three parts, making diagnosis for a certain different part only need to deal with the corresponding data set, effectively reducing the amount of data processing. Then this thesis puts forward a new organization discrimination method which is based on regional circularity, regional gray mean and variance. This thesis also proposes an entirely new improved active contour model, and combines it with three-dimensional region growing algorithm to make it a good solution to the head bone tissue extraction. A lot experimental comparison of clinical data has been done to prove that the proposed method is superior to the traditional threshold-based bone segmentation method. This thesis innovatively proposes a valley-like structure theory, and based on that theory further proposes an articular cartilage detection algorithm. By the articular cartilage detection algorithm, the separation of different bone issue can be effectively solved. In the final part, this thesis innovatively designs a new separate mandible structure extraction algorithm, and puts forward an independently designed mandible initial seed point quick determination algorithm and an independently designed articular cartilage detection algorithm. Through the two independently designed algorithms, the issue of separate mandible extraction is solved perfectly. The validity and accuracy of the new separate mandible structure extraction algorithm is verified by a large number of clinical data.