A Fast Subtraction Algorithm of Brain CT Angiography Images based on MMBE

Based on the traditional Matched Mask Bone Elimination (MMBE) method, this paper presents an improved subtraction method of Brain Computer Tomography Angiography (CTA) images. The method can be divided into two main parts, registration and bone elimination algorithms. Firstly, based on three-dimensional mutual information registration algorithm, this paper adapted sample and compression of gray scales through stastical analysis of the volume data to decrease data account and improve the calculate speed of registration. Then in the bone elimination algorithm, the extraction of bone mask in CT images is implemented by the combination of threshold and refinement of bone mask. We proposed a method to optimize the calculation strategies to decrease the cost of calculation. Through the analysis of the experiments on the clinical image data, the developed method proposed in this paper have improved about eight times of computing speed. The operation time of brain CT/CTA images with about two hundred slices has been limited in twenty seconds. And the main arteries can be displayed clearly in three-dimensional space.