Computational Zoom: A Framework for Post-Capture Image Composition (Supplementary Material)

2 DEFINING MULTI-PERSPECTIVE CAMERAS In this section, we describe in more detail how multiple pin-hole cameras along with a few 3D planes, which we call dolly planes, could be combined to form a single multi-perspective camera. To achieve this, multiple pin-hole cameras are rst transformed using plane-induced homographies before they are merged to de ne a multi-perspective camera. We will rst describe the e ect of applying the plane-induced homography to a pin-hole camera and why this operation is important in the context of computational zoom. Later, we will describe how to de ne a multi-perspective camera using multiple pin-hole cameras.