Texture Extraction and Blending without Prior Knowledge of Lighting Conditions

Authors:
H. L. Chou;C. C. Chen
Affiliations:
-;-
Venue:
PCM '02 Proceedings of the Third IEEE Pacific Rim Conference on Multimedia: Advances in Multimedia Information Processing
Year:
2002

Citing 9
Cited 0

A multiresolution spline with application to image mosaics

ACM Transactions on Graphics (TOG)
Detection of specularity using colour and multiple views

Image and Vision Computing - Special issue: 2nd European Conference on Computer Vision
Object shape and reflectance modeling from observation

Proceedings of the 24th annual conference on Computer graphics and interactive techniques
Rapid modeling of animated faces from video images

MULTIMEDIA '00 Proceedings of the eighth ACM international conference on Multimedia
3d Computer Graphics with Cdrom

3d Computer Graphics with Cdrom
Digital Image Processing

Digital Image Processing
Automatic Creation of 3D Facial Models

IEEE Computer Graphics and Applications
MPEG-4 Compatible Faces from Orthogonal Photos

CA '99 Proceedings of the Computer Animation
Texture Extraction from Photographs and Rendering with Dynamic Texture Mapping

ICIAP '99 Proceedings of the 10th International Conference on Image Analysis and Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Texture mapping techniques are widely used in photo-realistic 3D model rendering, but different lighting and viewing parameters create a difference in intensity for neighboring images, so that an edge appears at the boundary where neighboring images are stitched together. We propose an automatic procedure to extract and blend textures. Firstly, textures are extracted from the images and mapped to the triangles of the model. We choose the one with largest resolution if multiple textures extracted from different images are mapped to the same triangle. Secondly, a texture blending procedure is applied. We normalize the images to user-specified base images through overlapping area. Textures are then adjusted to the corresponding textures in the base images, if they exist. Finally, we check the boundary of neighboring textures. Boundary pixels varying discontinuously are averaged, and interior pixels are reassigned colors. Experimental results show that the smooth transition between neighboring textures provide better visual quality than just blending the boundary where neighboring images stitched.