基于形状保持主动轮廓的椭圆状目标检测(英文)

Elliptic Object Detection Based on Shape Preserving and Active Contour

本文通过形状约束方程(组)与一般主动轮廓模型结合,将目标形状与主动轮廓模型融合到统一能量泛函模型中,提出了一种形状保持主动轮廓模型即曲线在演化过程中保持为某一类特定形状。模型通过参数化水平集函数的零水平集控制演化曲线形状,不仅达到了分割即目标的目的,而且能够给出特定目标的定量描述。根据形状保持主动轮廓模型,建立了一个用于椭圆状目标检测的统一能量泛函模型,导出了相应的Euler-Lagrange常微分方程并用水平集方法实现了椭圆状目标检测。此模型可以应用于眼底乳头分割,虹膜检测及相机标定。实验结果表明,此模型不仅能够准确的检测出给定图像中的椭圆状目标,而且有很强的抗噪、抗变形及遮挡性能。

We integrate the detected object shape represented by shape restraint equation（s） and general active contour models into a unified energy function to generate the shape preserving active contour model, which keeps the curve to be a specific class shape in the evolvement. In this model, a specific class contour shape is represented as the zero level line of some certain parametric level set function（s）. So, this model can not only detect the given object correctly but also characterize the object shape quantificationally via these parameters. In addition, we build an energy functional for elliptic objects detection using the proposed model, deduce the corresponding Euler-Lagrange ordinary differential equation （ODE（s）） of its shape parameters and implement them using level sets method. The elliptic object detection model has many applications such as papilla segmentation, iris detection and camera calibration. By numerical experiments presented in the paper, it can not only detect the elliptic objects correctly but also is robust to noise, deformity and occlusion etc.