For the purpose of object boundary extraction, traditional shape-based deformable models rely on external image forces that come primarily from edge or image gradient information. Such reliance on local edge information makes the models prone to get stuck in local minima due to image noise and various other image artifacts. We propose a 2D deformable model—Metamorphs, which integrates region texture constraints so as to achieve more robust segmentation. Compared with traditional shape-based models, Metamorphs’s segmentation result is less dependent on model initialization and not sensitive to noise and spurious edges inside the object of interest. Then we present Active Volume Models (AVM), a similar and improved approach for 3D segmentation. The shape of this 3D model is considered as an elastic solid, with a simplex-mesh surface made of thousands of vertices. Deformations of the model are derived from a linear system that encodes external forces from the boundary of a Region of Interest (ROI), which is a binary mask representing the object region predicted by the current model. Efficient optimization and fast convergence of the model are achieved using the Finite Element Method (FEM). To further improve segmentation performance, a multiple-surface constraint is also employed to incorporate spatial constraints among multiple objects. It uses two surfacedistance based functions to adaptively adjust the weights of contribution from the image-based region information and from spatial constraints among multiple interacting surfaces. Several applications such as the rodent brain segmentation are shown to demonstrate the benefits of these segmentation algorithms based on deformable models that integrate multiple sources of constraints.