超肥胖人群的下肢体运动特征的图像分割仿真

Super Obese Body Movement Characteristics under the Simulation of Image Segmentation

针对超肥胖人群的下肢体运动特征的图像进行分割的过程中,由于超肥胖人员走路的特征较为特殊,不遵循正常的步态模式,需要加入很多特定的目标约束。传统的分割算法在图像中的信息量非常大时,受到这种特殊约束影响,造成分割耗时,效果差等问题。提出基于C-V和GCS的超肥胖人群的下肢体运动特征的图像分割模型,假设图像中所有同质区域的亮度均为常数,对能量函数进行处理,减少目标约束,从而使分割图像与原图像之间的差异达到最小。将边缘函数看作是权重函数对正则项进行改进和阈值化操作,获取特征函数。通过水平集法对超肥胖人群的下肢体运动特征的图像分割模型进行仿真。通过加性算子分裂法进行曲线仿真求和取平均值。仿真结果表明,所提方法的图像分割效果较优,同时具有较高的效率。

For super obese body motion characteristics in the process of image segmentation,due to the characteristics of the super obese individuals walk more special,do not follow the normal gait pattern,need to add a number of specific target constraint. Information in the traditional segmentation algorithms in the image is very large,affected by this particular constraints,causing segmentation time-consuming and poor effect. Put forward based on the C-V and GCS super obese body motion characteristics of the image segmentation model,assuming that all the brightness of the homogenous area in the image are constant,the energy function,reduces constraints,so that the image segmentation and the difference between the original image. Edge will function as a weighting function to improve the regularization and threshold operation,obtain the characteristic function. By the level set method for super obese body motion characteristics of the image segmentation model simulation. By the additive operator splitting method summation averaged curve simulation. The simulation results show that the proposed method of image segmentation effect is excellent,with high efficiency at the same time.