高级检索

几何活动轮廓模型的多尺度扩散分割算法

A Multi-Scale Diffusion Segmentation Algorithm Based on Geometric Active Contour Model

  • 摘要: 提出了一种对几何活动轮廓模型中的停止速度场进行多尺度扩散的算法,它通过引进2个控制参数来定义停止速度场的目标边界、同质区域和过渡区域;对于不同复杂性的图像,采取不同的控制参数对其停止速度场进行多尺度扩散;并将多尺度扩散后的停止速度场应用于几何活动轮廓模型进行图像分割.实验结果表明:对1幅合成图像和2幅自然图像,该算法大大地减少了分割时间,在一定程度上也减少了边界泄漏.

     

    Abstract: A mult-i scale diffusion algorithm is proposed for the halting speed fields (HSF) in geometrical active contours (GAC) model.By the algorithm,the object boundaries,homogenous regions and transition regions of the HSF are first defined through the introduction of two control parameters,and their values are set according to different complexities of the images.Then,the HSF is diffused by the mult-i scale diffusion algorithm and the diffused HSF is applied to the image segmentation based on GAC model.Experimental results on a synthetic image and two nature images demonstrate that the proposed algorithm can reduce both segmentation time and boundary leaking,in comparison with GAC model with the original HSF.

     

/

返回文章
返回