Depth Image Based 3D Object Model Generation

Abstract: This research studied the method of creating a complete 3D model by obtaining the location and r, g, b values of an object from the object's color and depth map images, and conducting the point based 3D modeling in OpenGL, and, finally, by filling the holes on the front or side according to the depth value. The most important aspect of the method is filling the front and sides of the object with differential hole-filling method. In addition, the 3D model could be viewed through 3D glasses by extracting left, right images from the 3D TV, and a view that corresponds to the location of the user could be supplied to the user by locating the user with an OptiTrack camera.