3D Indoor Scenes Modeling And Layout From A Single Image

The increasing digital media are appearing in public views more frequently and widely nowadays. A great many of pictures and models are penetrating into many field-s. However, how to reconstruct a 3D virtual scene from the image is an unavoidable problem when people trying to describe the real world. And it, therefore, becomes a hot field in computer graphics. To assist the rapid scene design, we present a bran-new approach based on the previous research. Users can use our method reconstruct the 3D indoor scene based on a single image rapidly, and then do the scene layout. Further-more, we develop a prototype system based on those research. Generally, our paper includes the following three main contributions:(1) 3D Indoor Scenes Modeling from a Single Image. With a single input indoor image, a 3D scene can be reconstructed from a single image using an existing model library in three stages:image analysis, model retrieval and relevance feedback. In the image analysis stage, we obtain the object information from the input image using geometric reasoning technology combined with an image segmentation method. In the model retrieval stage, line drawings are extracted from 2D objects and 3D models by using different line rendering methods. We exploit various tokens to represent local features and then organize them together as a star-graph to show a global description. By comparing similarity among the encoded line drawings, models are retrieved from the model library. Also, for a better user experience, we add a relevance feedback stage following the retrieval stage. The Support Vector Machine method is used to conduct the feedback operation. After this stage, the retrieved models are in conformance with the image semantic. The 3D scene is then reconstructed. Experimental results show that driven by the given model library, indoor scenes modeling from a single image could be achieved automatically and efficiently.(2) An Indoor Scenes Layout Method. We present an indoor scenes layout method following the scene modeling approach that assists users to do a better de-sign. Our approach summarizes guidelines presented and divide them into two parts: functional guidelines and visual guidelines. These guidelines are constraints when do-ing the scene layout. In the procedure of furniture layout, we use the approach based on the Markov chain Monte Carlo sampler, combine the constraints into density terms to guide the layout suggestions and generate several suggested scenes from the sampler. Our experiments show that assisted with a little interaction, our method gives a good guidance to users for their indoor scene layout. We accelerate the problem of scene layout in an efficient way.(3) Design and Implementation of IndoorDesign3D System. We reasonably combine the previous researches into the prototype system of IndoorDesign3D. For the perspective of software design, we introduce the system functionality, data structure, and the main class diagrams. And we illustrate the main modules of our system. Above all, our system unifies the scene reconstruction and layout approach, providing users with a more efficient and convenient experience in the scene design.