Depth map estimation from single-view image using object classification based on Bayesian learning

Generation of three-dimensional (3D) scenes from two-dimensional (2D) images is an important step for a successful introduction to 3D multimedia services. Among the relevant problems, depth estimation from a single-view image is probably the most difficult and challenging task. In this paper, we propose a new depth estimation method using object classification based on the Bayesian learning algorithm. Using training data of six attributes, we categorize objects in the single-view image into four different types. According to the type, we assign a relative depth value to each object and generate a simple 3D model. Experimental results show that the proposed method estimates depth information properly and generates a good 3D model.