Shadow generation of linear light source without 3D reconstruction

  • Authors:
  • Jen-Hui Chuang;Jau-Hong Kao

  • Affiliations:
  • Department of Computer Science, National Chiao Tung University, Hsinchu City, Taiwan;Department of Computer Science, National Chiao Tung University, Hsinchu City, Taiwan

  • Venue:
  • ICCOMP'06 Proceedings of the 10th WSEAS international conference on Computers
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents a novel image-based approach to generating object shadows. Previous approaches of shadow generation are usually based on 3D geometry. In this paper, we present a method of deriving accurate shadows of a 3D object using multiple images without 3D reconstruction. For each view of the scene, the object shadows due to a linear light source connecting the two view points can be generated. The shadow generation is accomplished by identifying shadow region in an image with respect to some reference points on a base plane. The proposed approach uses view-invariant cross ratio and some geometric properties associated with perspective projection. It is also extended to generate shadows for more general scenes where multiple objects and multiple projection planes on which object shadows are to be cast are involved. Experiments show that the shadows generated for both real and synthetic scenes are satisfactory and the computation is efficient.