3D modeling is an important and interesting problem in both computer vision and computer graphics fields. Recently, image based modeling is receiving more research focuses because it is of low cost, easy to handle and can generate models with very high precision. The technology is also widely used in digital cultural heritage, movies, games, video surveillance, etc. This technology is highly valuable in both research and application. Image based modeling focuses on reconstructing 3D models of objects or scenes directly from single image, image sequences or videos. The whole modeling procedure can be fully automated or facilitated by human interaction. The key problem of image based modeling is shape from image, which focuses on reconstructing 3D geometry information from images. However, review work is not available on this topic recently, which has obviously blocked the development of image based modeling research. As a result, presented in this paper is a review of analysis and discussion on shape from image. Different methods are classified based on the visual clues used in modeling from a computer vision perspective. Basic functions and related works are also introduced for each method. After analyzing and comparing these methods individually, a conclusion is drawn which includes the characteristics, problems and future of shape from image.