In this paper, image segmentation based on an integrated vector-valued active contour model is presented. Consider that each image channel has its signal characteristics, the region-based information uses the hybrid mean intensities simultaneously. Furthermore, by utilizing a two-dimensional vector field with different image channels, which provides different image patterns are used to constrain the results of image segmentation using edge-based information, edge-based information is also used, which extracts the information and uses the nonlinear function to learn the specific segmentation process. With the incorporation of the vector-based region and edge information, the proposed method can deal with multi-channel images effectively. The method is applicable for color images, multiresolution representation from frequency transformation, and multi-modal images, and can effectively overcome the problem that the weak edge of the image cannot converge due to large noise and poor contrast. It can confirm the effectiveness and robustness of the proposed method.

在本文中，提出了基于集成向量值活动轮廓模型的图像分割。考虑到每个图像通道都有其信号特征，基于区域的信息同时使用混合平均强度。此外，通过利用具有不同图像通道的二维矢量场，提供不同的图像模式来约束使用基于边缘信息的图像分割结果，还使用基于边缘信息，提取信息并使用非线性函数来学习具体的分割过程。通过结合基于矢量的区域和边缘信息，该方法可以有效地处理多通道图像。该方法适用于彩色图像、频率变换的多分辨率表示、和多模态图像，可以有效克服图像弱边缘因噪声大、对比度差而无法收敛的问题。它可以证实所提出方法的有效性和鲁棒性。