Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network

  Advanced Search  

Multiscale Shape and Detail Enhancement from Multi-light Image Collections Raanan Fattal

Summary: Multiscale Shape and Detail Enhancement from Multi-light Image Collections
Raanan Fattal
University of California, Berkeley
Maneesh Agrawala
University of California, Berkeley
Szymon Rusinkiewicz
Princeton University
Input: 3 MLIC Images Our Results: Enhanced Shape and Surface Detail
Figure 1: The Multi-Light Image Collection for this chard leaf contains 3 images taken under varying lighting conditions. The shading in each input image
reveals different aspects of its shape and surface details. We combine the shading at multiple scales across the input images to generate the enhanced results.
The result on the left exaggerates surface details by eliminating shadows, but yields a flat look. The result on the right is less extreme and includes some
shadows to increase the perception of depth, at the cost of reducing some visible detail in the shadow regions.
We present a new image-based technique for enhancing the shape
and surface details of an object. The input to our system is a small
set of photographs taken from a fixed viewpoint, but under varying
lighting conditions. For each image we compute a multiscale de-
composition based on the bilateral filter and then reconstruct an en-
hanced image that combines detail information at each scale across
all the input images. Our approach does not require any informa-


Source: Agrawala, Maneesh - Department of Electrical Engineering and Computer Sciences, University of California at Berkeley
Princeton University, Department of Computer Science, Princeton Graphic Group


Collections: Computer Technologies and Information Sciences