“…A UTOMATIC 3D geometric labeling or layout reasoning from a single scene image is one of the most important and challenging problems in scene understanding. It offers the mid-level information for other high-level scene understanding tasks such as 3D world reconstruction [1], [2], [3], [4], [5], [6], depth map estimation [7], [8], [9], scene classification [10], [11], [12] and content-based image retrieval [13], [14], [15].…”