Plan De Cuentas

An Augmented Reality Application for Previewing 3D Décor Changes
Mai Le, Aaron Zarraga, Kangrong Zhu
Department of Electrical Engineering Stanford University, CA, USA a view matrix can be calculated. The perspective projection matrix reflects the transform a 3D object undergoes under a perspective projection, and the view matrix reflects the transform a 3D object undergoes due to differentpositions and orientations of the camera. OpenGL can use the perspective projection matrix and the view matrix as input and render a 3D model into a scene with correct perspective projection and object position so that the object conforms to the perspective lines present in the scene. Extensions for Android Application To expand upon the interior design applications of this project, we also exploredmethods for wall detection and coloring. However, since this functionality is not central to the project, we simply demonstrate a proof of concept, rather than a completely developed method, for a wall painting algorithm. II. RELATED WORK Previous work has used augmented-reality to preview 2D planar object [2]. That work was based on the assumption that a marker undergoes an affine transform whenan image is taken. However, this assumption fails when the object changes from 2D to 3D. Instead of an affine transform, a perspective transform should be applied on the marker. In our work, we estimate the parameters for a perspective transform based on the marker inside an image, and render a 3D object into a captured scene using OpenGL. To calculate the projection matrix and the view matrix forthe perspective transform, we need to calculate two vanishing points for the two orthogonal directions. Although marker-less vanishing point calculation algorithm exists [1], we chose a marker-based approach for two reasons: (1) many parameters for correctly rendering the 3D model, for example the origin to place the model, the reference RGB value for color balancing, and the parameters for modelorientation adjustment , could be calculated based on a carefully designed marker. (2) A marker based approach is more robust and easier to implement than a marker-less approach. MARKER DESIGN AND VANISHING POINT STRATEGY We first considered using a characteristic pattern on our marker and a feature detection algorithm such as SIFT to locate the marker within the image. Because feature detectionis slow, we decided to try a faster, contrast-based detection method, and III. C.

Abstract—In this paper, we explore a marker-based method for placement of 3D objects in an indoor scene for the purpose of previewing décor changes. Objects are placed in the scene using OpenGL using a MATLAB vanishing point calculation algorithm. This technique allows users to specify orientation and placement ofthe object via placement of a physical marker and allows for color balance correction on the 3-D object. We also conducted preliminary exploration into wall detection and coloring techniques. Keywords- augmented reality; marker detection; vanishing point; 3D model; mobile application

I.

INTRODUCTION

A. Motivation Online retailers have historically been unable to compete with localretailers in their inability to allow users to visualize the product in person. Clothing retailers such as Lands’ End have created customizable virtual models so that online shoppers can visualize the fit and size of clothing items. However, customizable models are not feasible for the simulation of decor changes and furniture additions, due to the variation in room types, sizes, and existing furnitureconfigurations. Subtleties such as color matching to existing fixtures are also difficult to replicate in a simulated environment. For this reason, an augmented reality mobile application is an ideal candidate for allowing customers to visualize the addition of a furniture element or decor changes such as wall painting. The use of easy to print paper markers would allow for easy use in the home...