Markov random field models in computer vision
ECCV '94 Proceedings of the third European conference on Computer Vision (Vol. II)
Image quilting for texture synthesis and transfer
Proceedings of the 28th annual conference on Computer graphics and interactive techniques
Efficient Graph-Based Image Segmentation
International Journal of Computer Vision
Optimizing the parameters for patch-based texture synthesis
Proceedings of the 2006 ACM international conference on Virtual reality continuum and its applications
| Hi-index | 0.00 |
Patch-based texture synthesis method uses MRF texture model to synthesize a bigger texture from a smaller patch sample containing two user-defined parameters, patch size and boundary zone. To obtain optimal values for the parameters, the texture has to be analyzed, which costs too expensive for real-time large texture synthesis. This paper introduces a more efficient method for finding the optimal value of the two parameters. Firstly, we use graph-based image segmentation to extract feature segments from the input sample. We then choose a set of major segments preserving the main features to appear in the final result. Finally, we calculate the two parameters based on size and repetition of the segments. The experimental results how that our technique can reduce computational time for determining the parameters compared to previous method and can work with several type of textures.