Lattice Boltzmann Model based 3D Image Segmentation

2nd International Symposium on Application of Materials Science and Energy Materials

材料科学;能源学

Chen, Yu^1 ; Lu, Dongxiang^1

School of Information Engineering, Yancheng Teachers University, Yancheng, Jiangsu, China^1

关键词: 3D image segmentation;    Curve evolution methods;    Geometric active contours;    Lattice boltzmann models;    Natural parallelism;    Three dimensional images;    Three dimensional segmentation;    Three-dimensional volume;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/490/7/072005/pdf
DOI  :  10.1088/1757-899X/490/7/072005

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

Image segmentation plays a very important role in three-dimensional volume data processing. In three-dimensional segmentation, the curve evolution method based on geometric active contour model and level set method is one of the most widely concerned methods. Among these methods, the CV model is a hot research topic. Based on the Munfor-Shah model, CV model is proposed by Chan and Vese. Because of the huge amount of data in three-dimensional images, the computation of solving CV model is very large, which consumes a lot of time. In this study, a 3D image segmentation algorithm based on Lattice Boltzmann model is proposed. The experiment shows our method is effective. The algorithm is simple and efficient. More importantly, the LB model has natural parallelism. Especially, our method can be implemented on massively parallel image processing platform, for example, FPGAs, DSPs and GPUs.