In the field of traditional computer vision, the theory, in which the visual interpretation task comprises several levels that can be managed independently, has great influence on researchers. It presents that the information for accomplishing low level visual task can only be obtained from image itself. Kass et al. challenged the theory by developing an active contour model called Snake in 1987. Since then, this model has been enjoying a wide range of applications in the field of computer vision and significant advances have been made. The paper reviews the research, development and applications of the active contour model, and presents possible future research orientations.