Improved methods of estimating shape from shading using the light source coordinate system
Artificial Intelligence
A Method for Enforcing Integrability in Shape from Shading Algorithms
IEEE Transactions on Pattern Analysis and Machine Intelligence
Face Recognition: The Problem of Compensating for Changes in Illumination Direction
IEEE Transactions on Pattern Analysis and Machine Intelligence
Acquiring the reflectance field of a human face
Proceedings of the 27th annual conference on Computer graphics and interactive techniques
The Quotient Image: Class-Based Re-Rendering and Recognition with Varying Illuminations
IEEE Transactions on Pattern Analysis and Machine Intelligence
From Few to Many: Illumination Cone Models for Face Recognition under Variable Lighting and Pose
IEEE Transactions on Pattern Analysis and Machine Intelligence
Proceedings of the 28th annual conference on Computer graphics and interactive techniques
Reflectance from Images: A Model-Based Approach for Human Faces
IEEE Transactions on Visualization and Computer Graphics
CVPR '05 Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) - Volume 2 - Volume 02
Post-production facial performance relighting using reflectance transfer
ACM SIGGRAPH 2007 papers
Fourier Analysis of the 2D Screened Poisson Equation for Gradient Domain Problems
ECCV '08 Proceedings of the 10th European Conference on Computer Vision: Part II
Smooth reconstruction and compact representation of reflectance functions for image-based relighting
EGSR'04 Proceedings of the Fifteenth Eurographics conference on Rendering Techniques
EGSR'05 Proceedings of the Sixteenth Eurographics conference on Rendering Techniques
| Hi-index | 0.00 |
We propose and analyze an example-based framework for relighting images. In this framework, there are a number of images of reference objects captured under different illumination conditions. Given an input image of a new object captured under one of the previously observed illumination conditions, new images can be synthesized for the input object under all the other illumination conditions that are present in the reference images. It does not require any other prior knowledge on the reference and target objects, except that they share the same albedo. Though it is appreciated if the reference objects have similar shape as the target object, sphere or ellipsoid which has plenty of local geometry samples are sufficient to build up a look-up table, as this method solves the problem locally. Gradient domain methods are introduced to finally generate visual-pleasing results. We demonstrate this framework on synthesized data and real images.