A morphable model for the synthesis of 3D faces
Proceedings of the 26th 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
Lambertian Reflectance and Linear Subspaces
IEEE Transactions on Pattern Analysis and Machine Intelligence
Multilinear Analysis of Image Ensembles: TensorFaces
ECCV '02 Proceedings of the 7th European Conference on Computer Vision-Part I
Face Recognition in Hyperspectral Images
IEEE Transactions on Pattern Analysis and Machine Intelligence
A Bilinear Illumination Model for Robust Face Recognition
ICCV '05 Proceedings of the Tenth IEEE International Conference on Computer Vision - Volume 2
Multi-Modal Tensor Face for Simultaneous Super-Resolution and Recognition
ICCV '05 Proceedings of the Tenth IEEE International Conference on Computer Vision - Volume 2
A Subspace Model-Based Approach to Face Relighting Under Unknown Lighting and Poses
IEEE Transactions on Image Processing
Hi-index | 0.00 |
In this paper, a new approach to face relighting by the product of reflectance image and illumination Tensorfaces is proposed With a pair of multi-spectral images, a near infrared and a visual image, the intrinsic images decomposition can be implemented and corresponding reflectance image is derived Besides, the illumination images obtained from last step as well as the input visual images constitute a 3-D tensor, on which super-resolution and maximum a posteriori probability estimation are carried out And then, illumination Tensorfaces under specific light are derived, by which face under target illumination can be synthesized In contrast to commonly used shape models or shape dependent models, the proposed method only relies on Lambertian assumption and manages to recover reflectance of the face Besides, compared with the existing methods, i.e Tensorfaces and Quotient Image, our methods properly preserve the identity of the subject as well as the texture details Experiments show that the proposed method is not only simple when deriving intrinsic images, but also practical when performing face relighting.