高强度运动下骨骼CT图像骨裂纹智能检测仿真

发布时间：2018-02-11 07:28

本文关键词： 高强度运动图像骨裂缝检测　出处：《计算机仿真》2017年05期 　论文类型：期刊论文

【摘要】：高强度运动下对骨骼CT图像骨裂纹的智能检测,可以有效地预防高强度运动下对骨骼的损伤。对CT图像骨裂纹进行智能检测,需要先将上述问题转换为对偶问题,获取CT图像纹理骨裂缝最优分类函数,完成对图像骨裂纹的检测。传统方法给出图像的边界的直径,计算出图像的长轴与短轴的比值,但忽略了对骨裂缝纹理的分类,导致检测精度低。为此提出基于GLCM的骨骼CT图像骨裂纹智能检测方法。基于统计思想组建骨骼CT图像灰度共生矩阵,选择不同的纹理骨裂缝参数值表述CT图像的纹理骨裂缝,融合于支持向量机理论对获取的CT图像的纹理骨裂缝进行分类,引入Lagrange乘子将对骨骼CT图像骨裂缝分类问题转换为对偶问题,获取CT图像纹理骨裂缝最优分类函数。利用上述函数完成对骨骼CT图像骨裂纹智能检测。实验结果表明,所提方法智能检测精确度高,并且可以为医生诊断提供科学的依据。
[Abstract]:The intelligent detection of bone cracks in bone CT images under high intensity exercise can effectively prevent bone damage under high intensity exercise. In order to detect bone cracks intelligently in CT images, the above problems need to be converted into dual problems. The optimal classification function of bone fracture in CT image is obtained, and the detection of bone crack is completed. The traditional method gives the diameter of the boundary of the image, calculates the ratio of the long axis to the short axis of the image, but neglects the classification of the bone crack texture. Therefore, an intelligent detection method of bone crack in bone CT image based on GLCM is proposed. Based on statistical idea, gray level co-occurrence matrix of bone CT image is constructed, and different texture bone crack parameter values are selected to express the texture bone crack in CT image. Based on the theory of support vector machine (SVM), the texture fractures of CT images are classified, and the Lagrange multiplier is introduced to transform the classification of bone fractures into dual problems. The optimal classification function of texture bone cracks in CT images is obtained. The intelligent detection of bone cracks in bone CT images is accomplished by using the above functions. The experimental results show that the proposed method has high accuracy and can provide scientific basis for doctors' diagnosis.
【作者单位】：郑州大学体育学院;
【基金】：河南省哲学社会科学规划项目(2015BZZ009)
【分类号】：R873;TP391.41

【参考文献】