Abstract: Devices and techniques are described for augmented reality furniture layout recommendation. In some examples, first data comprising a three dimensional (3D) model of a room is received. Second data representing an intended usage of the room may be received. A first machine learning model may generate a first collection template using the first data and the second data. The first collection template may include first bounding box data representing a first category of furniture and second bounding box data representing a second category of furniture. The first bounding box data representing a first bounding box having a first size and located at a first position within the 3D model at a first orientation. The second bounding box data representing a second bounding box having a second size and located at a second position within the 3D model at a second orientation.

Abstract: A machine learning system receives a reference image and generates a series of projected view images of a physical object represented in the images. Parallel neural networks may receive the reference image and series of projected view images for analysis to determine one or more features of the physical object. By pooling the results from the parallel network, a single output may be provided to a set of decodes that are trained to identify a material property of the one or more items. As a result, a three-dimensional model may be generated that includes a graphical representation of the object as a function of its material properties to enable improved rendering.