Synthesizing natural textures
ACM Symposium on interactive 3D graphics, pp.217-226, (2001)
We present a simple texture synthesis algorithm that is well-suited for a specific class of naturally occurring textures. This class in- cludes quasi-repeating patterns consisting of small objects of fa- miliar but irregular size, such as flower fields, pebbles, forest un- dergrowth, bushes and tree branches. The algorithm starts from a s...更多
下载 PDF 全文
- Texture mapping  is a common method that adds realism to computer generated images.
- Dischler and Ghazanfarpour  address a more difficult problem of extracting 3D macrostructures from texture images
- Once this is done, the shape of the macrostructures can be interactively changed and mapped onto a model.
- The shape of the macrostructures can be interactively changed and mapped onto a model
- Since this problem is much more complicated than the usual 2D texture synthesis, it is not surprising that their algorithm is much more involved than the one the authors present
- Texture mapping  is a common method that adds realism to computer generated images
- Most texture synthesis algorithms attempt to deal with as large class of textures as possible
- The particular class of textures this paper deals with includes many naturally occuring quasi-repeating patterns consisting of familiar elements, such as pebbles, leaves, bushes, flowers, tree branches, etc
- Easy to implement, efficient and produces good visual results. These qualities allow it to be extended to handle direct user interaction with the synthesis process and create textures based on a user-created input image
- The technique used to take into account user input is not unique to our algorithm and can probably be used with the original WL algorithm once processor speeds make the iteration of the WL algorithm faster
- Results and Discussion
Some synthesis results are shown on Figure 9 along with WLgenerated images.
- Information necessary to find pixels near boundaries between individual texture patches is readily available in the algorithm (Figure 5) and, if desired, some selective blurring can be applied to these pixels to help hide the seams.
- This has not been done for any of the images presented in this paper
- The WL algorithm works surprisingly well given its simplicity. For many textures the quality of synthesized images is as good as with any other known technique but with runtime orders of magnitude faster.
- A black-and-white image is the extreme case of this, see Figure 3, top left
- Another case where applying the algorithm produces excellent results occurs when the underlying textures are smooth and the absence of well-defined edges is what is wanted.
- Easy to implement, efficient and produces good visual results
- These qualities allow it to be extended to handle direct user interaction with the synthesis process and create textures based on a user-created input image.
- The technique used to take into account user input is not unique to the algorithm and can probably be used with the original WL algorithm once processor speeds make the iteration of the WL algorithm faster
- This work was supported by NSF awards 97-31859, 97-20192 and 8920219
- Vision texture library. MIT Media Lab. http://wwwwhite.media.mit.edu/vismod/imagery/VisionTexture/.
- BLINN, J., AND NEWELL, M. Texture and reflection in computer generated images. Communications of the ACM 19, 10 (October 1976), 542–547.
- BOLIN, M. R., AND MEYER, G. W. A perceptually based adaptive sampling algorithm. Proceedings of SIGGRAPH 98 (July 1998), 299–309.
- DANA, K., VAN GINNEKEN, B., NAYAR, S., AND KOENDERINK, J. Reflectance and texture of real world surfaces. ACM Transactions on Graphics 18, 1 (January 1999), 1–34.
- DEBONET, J. S. Multiresolution sampling procedure for analysis and synthesis of texture images. Proceedings of SIGGRAPH 97 (August 1997), 361–368.
- DISCHLER, J.-M., AND GHAZANFARPOUR, D. Interactive image-based modeling of macrostructured textures. IEEE Computer Graphics and Applications 19, 1 (JanuaryFebruary 1999), 66–74.
- EFROS, A., AND LEUNG, T. Texture synthesis by nonparametric sampling. International Conference on Computer Vision 2 (September 1999), 1033–1038.
- HARALICK, R. Statistical image texture analysis. In Handbook of Pattern Recognition and Image Processing, vol.
- 86. Academic Press, June 1986, pp. 247–279.
-  HEEGER, D. J., AND BERGEN, J. R. Pyramid-based texture analysis/synthesis. In Proceedings of SIGGRAPH 95 (Anaheim, California, August 6–11, 1995) (August 1995), R. Cook, Ed., Computer Graphics Proceedings, Annual Conference Series, pp. 229–238.
-  MIYATA, K. A method of generating stone wall patterns. Computer Graphics 24, 3 (August 1990), 387–394. ACM Siggraph ’90 Conference Proceedings.
-  PERLIN, K. An image synthesizer. Computer Graphics 19, 3 (July 1985), 287–296. ACM SIGGRAPH 85 Conference Proceedings.
-  PERLIN, K., AND HOFFERT, E. M. Hypertexture. Computer Graphics 23, 3 (July 1989), 253–262. ACM SIGGRAPH 89 Conference Proceedings.
-  PORTILLA, J., AND SIMONCELLI, E. P. A parametric texture model based on joint statistics of complex wavelet coefficients. International Journal of Computer Vision 39, 3 (October 2000). to appear.
-  PRAUN, E., FINKELSTEIN, A., AND HOPPE, H. Lapped textures. Proceedings of SIGGRAPH 2000 (July 2000), 465– 470.
-  RAMASUBRAMANIAN, M., PATTANAIK, S. N., AND GREENBERG, D. P. Perceptually based physical error metric for realistic image synthesis. Proceedings of SIGGRAPH 99 (August 1999), 73–82.
-  WEI, L.-Y., AND LEVOY, M. Fast texture synthesis using tree-structured vector quantization. Proceedings of SIGGRAPH 2000 (July 2000), 479–488.
-  WITKIN, A., AND KASS, M. Reaction-diffusion textures. In Computer Graphics (SIGGRAPH ’91 Proceedings) (July 1991), T. W. Sederberg, Ed., vol. 25, pp. 299–308.
-  XU, Y., GUO, B., AND SHUM, H.-Y. Chaos mosaic: Fast and memory efficient texture synthesis. Tech. Rep. MSR-TR2000-32, Microsoft Research, 2000.