A geometric model for active contours in image processing is proposed, based on a geometric partial differential equation. The model is intrinsic, stable (satisfies the maximum principle), and allows rigorous mathematical analysis. It enables the extraction of smooth shapes (angles cannot be retrieved) and can be adapted to find multiple contours simultaneously. The stability allows for the design of robust algorithms with no parameters. Numerical experiments are presented. The paper discusses two methods for edge detection: the classical snake method and a new geometric PDE model based on mean curvature motion. The snake method starts with an initial contour and seeks deformations that move it toward the desired contour by minimizing an energy functional. The energy functional consists of internal energy (for smoothness) and external energy (for image features). The internal energy is defined as the integral of the curve's curvature and torsion, while the external energy is defined as the integral of the image gradient. The snake method aims to minimize the total energy to find edges. The snake method provides a global view of edge detection, differing from traditional approaches that detect edges and then link them. It has been developed by several researchers, including Kass-Witkin-Terzopoulos, Blake-Zisserman, and others. The method is a way of regularizing the edge detection problem, which is ill-posed. The paper compares the snake method with the new geometric PDE model and discusses their strengths and weaknesses in addressing the issues of edge detection.A geometric model for active contours in image processing is proposed, based on a geometric partial differential equation. The model is intrinsic, stable (satisfies the maximum principle), and allows rigorous mathematical analysis. It enables the extraction of smooth shapes (angles cannot be retrieved) and can be adapted to find multiple contours simultaneously. The stability allows for the design of robust algorithms with no parameters. Numerical experiments are presented. The paper discusses two methods for edge detection: the classical snake method and a new geometric PDE model based on mean curvature motion. The snake method starts with an initial contour and seeks deformations that move it toward the desired contour by minimizing an energy functional. The energy functional consists of internal energy (for smoothness) and external energy (for image features). The internal energy is defined as the integral of the curve's curvature and torsion, while the external energy is defined as the integral of the image gradient. The snake method aims to minimize the total energy to find edges. The snake method provides a global view of edge detection, differing from traditional approaches that detect edges and then link them. It has been developed by several researchers, including Kass-Witkin-Terzopoulos, Blake-Zisserman, and others. The method is a way of regularizing the edge detection problem, which is ill-posed. The paper compares the snake method with the new geometric PDE model and discusses their strengths and weaknesses in addressing the issues of edge detection.