• Abstract: Given a single photo of a room and a large database of furniture CAD models, our goal is to reconstruct a scene that is as similar as possible to the scene depicted in the photograph, and composed of objects drawn from the database.
• By operating jointly on the full scene at once, we account for inter-object occlusions.
• Our approach iteratively optimizes the placement and scale of objects in the room to best match scene renderings to the input photo, used image comparison metrics trained using deep convolutional neural nets.
• We present a completely automatic system to address this IM2CAD problem that produces high quality results on challenging imagery from real estate web sites.
• Computer Vision and Pattern Recognition (cs.CV)

Read more