|
Title:
|
RENDERING FROM UNSTRUCTURED IMAGE COLLECTIONS |
|
Author(s):
|
Siren Bato , Yoshinori Dobashi , Tsuyoshi Yamamoto |
|
ISBN:
|
978-972-8924-97-3 |
|
Editors:
|
Hans Weghorn and Pedro Isaías |
|
Year:
|
2009 |
|
Edition:
|
V I, 2 |
|
Keywords:
|
Image-based rendering, computer vision, virtual reality. |
|
Type:
|
Full Paper |
|
First Page:
|
369 |
|
Last Page:
|
378 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Many image based rendering (IBR) techniques have been developed to render new viewpoint images from multiple precaptured
images. This paper presents a new IBR method for synthesizing new viewpoint images from a large
unstructured image collection, i.e., the input images are randomly distributed and not restricted to a plane or to any
specific manifold. Unlike the methods that sample, construct and re-sample the Plenoptic function, in the proposed
method, the rendering problem is formulated as an optimization problem of finding the optimal pixel from the input
images. The scene structure information (approximate geometry models) is used as additional input to render novel
views. Our approach makes it possible to build a large scale indoor or outdoor scene from a reasonable number of input
images. Different from most of the existing image based rendering systems, the proposed system is compatible with the
standard model based rendering systems, which allows synthesizing the target scene with other virtual objects or scenes. |
|
|
|
|
|
|
