Academia.eduAcademia.edu

Outline

Title
Abstract
Figures
Introduction
Previous Work
Overview
Finding the Best Patch
Results
Conclusions and Future Work
References
All Topics
Computer Science
Computer Graphics

Hybrid Texture Synthesis

Profile image of Andy NealenAndy Nealen

2003, … of the 14th Eurographics workshop on …

Abstract
sparkles

AI

The paper presents a novel adaptive hybrid texture synthesis method aimed at generating high-quality textures by integrating various successful techniques. By employing a patch-based synthesis framework with enhancements such as overlap re-synthesis and adaptive patch sampling, the method overcomes limitations of previous approaches, particularly in preserving detailed and global structures. Experimental results demonstrate significant improvements in texture synthesis quality, showcasing the effectiveness of the proposed algorithm in addressing common artifacts found in existing methods.

Figures (13)
‘e 1: In algorithm step i (with the black area in (a) to be synthesized) we select a new patch from the input texture, constrained by the overlap with the already existing synthesis result Rj-, (a). Here, the algorithm finds a good match (with A; < Amax) and does not adaptively split the patch. Pixels in the overlap region, which exceed a user- defined pixel error tolerance (Smax € (0, 1]) are marked as invalid (white) and the new patch P; is composited into R;_- (b). Then the invalid pixels are re-synthesized (c). The final patch P; is constrained on its entire boundary, ensuring that the resulting texture will tile seamlessly (d). max = 0.03, Amax = 0.05, ov = 6, initial patch size 32x32.
‘e 1: In algorithm step i (with the black area in (a) to be synthesized) we select a new patch from the input texture, constrained by the overlap with the already existing synthesis result Rj-, (a). Here, the algorithm finds a good match (with A; < Amax) and does not adaptively split the patch. Pixels in the overlap region, which exceed a user- defined pixel error tolerance (Smax € (0, 1]) are marked as invalid (white) and the new patch P; is composited into R;_- (b). Then the invalid pixels are re-synthesized (c). The final patch P; is constrained on its entire boundary, ensuring that the resulting texture will tile seamlessly (d). max = 0.03, Amax = 0.05, ov = 6, initial patch size 32x32.
Table 1: List of used variables/terms 1.2. Overview
Table 1: List of used variables/terms 1.2. Overview
Figure 4: The input texture T (left), the binary support func- tion J; (middle) and the image mask I; for the example out- lined in Figure 1 The image mask /; and binary support function J; (Fig. 4) are needed for the error image computation (Fig. 5). BUILDIMAGEMASK simply grows the current patch by the pixel overlap ov and checks the grown region in the current result R;_, for already synthesized pixels: if there exists a valid pixel, J; (initially all 0’s) is set to 1 and the color value from R;_ 1 (Fig. 1(a)) is stored in J;. Figures 1(a) and 4 show that for this patch we must match on its entire boundary to produce a tileable texture.
Figure 4: The input texture T (left), the binary support func- tion J; (middle) and the image mask I; for the example out- lined in Figure 1 The image mask /; and binary support function J; (Fig. 4) are needed for the error image computation (Fig. 5). BUILDIMAGEMASK simply grows the current patch by the pixel overlap ov and checks the grown region in the current result R;_, for already synthesized pixels: if there exists a valid pixel, J; (initially all 0’s) is set to 1 and the color value from R;_ 1 (Fig. 1(a)) is stored in J;. Figures 1(a) and 4 show that for this patch we must match on its entire boundary to produce a tileable texture.
Figure 5: The error Image E; (left), the trimmed error im- age Ej trim with the selected patch P; superimposed (middle, the overlap region is darkened), and P; with I; from Fig- ure 4 superimposed to highlight the remaining artifacts be- fore overlap re-synthesis (right). See Figure 12 for a com- parison of overlap repair strategies using this example. model !!. This accounts for the greater sensitivity of the hu- man visual system to changes in luminance than changes in hue or saturation !!. Note that this is not equivalent to match- ing on gray levels.
Figure 5: The error Image E; (left), the trimmed error im- age Ej trim with the selected patch P; superimposed (middle, the overlap region is darkened), and P; with I; from Fig- ure 4 superimposed to highlight the remaining artifacts be- fore overlap re-synthesis (right). See Figure 12 for a com- parison of overlap repair strategies using this example. model !!. This accounts for the greater sensitivity of the hu- man visual system to changes in luminance than changes in hue or saturation !!. Note that this is not equivalent to match- ing on gray levels.
Figure 7: Balancing Amax. Left: Amax = 0.01, resulting in many small patches, comparable to pixel-based algorithms. Middle: Amax = 1.0, which uses only the initial, uniform patch list and fails to eliminate all overlap artifacts. Right: Amax = 0.04, resulting in a good trade-off between global and local structure. The images are all 128X128 results oj Figure 13(a) Figure 6: Adaptive patch sampling. Shades of gray repre- sent already synthesized patches, the hatched areas are the overlap regions used for best patch search.
Figure 7: Balancing Amax. Left: Amax = 0.01, resulting in many small patches, comparable to pixel-based algorithms. Middle: Amax = 1.0, which uses only the initial, uniform patch list and fails to eliminate all overlap artifacts. Right: Amax = 0.04, resulting in a good trade-off between global and local structure. The images are all 128X128 results oj Figure 13(a) Figure 6: Adaptive patch sampling. Shades of gray repre- sent already synthesized patches, the hatched areas are the overlap regions used for best patch search.
in HYBRIDSYNTHESIZE and Figure 6). This leads to adap- tive sampling grids such as the one shown in Figure 6, bot- tom right, and Figure 15, right column. Note that we halve the pixel overlap ov for recursive calls, choosing a minimal value of ov = 3 (line 12 in HYBRIDS YNTHESIZE).
in HYBRIDSYNTHESIZE and Figure 6). This leads to adap- tive sampling grids such as the one shown in Figure 6, bot- tom right, and Figure 15, right column. Note that we halve the pixel overlap ov for recursive calls, choosing a minimal value of ov = 3 (line 12 in HYBRIDS YNTHESIZE).
Figure 8: The error in the overlap region Sj, defined ir Equation 3 (left; the error-values S;(x) are normalized to the interval [0,1] for visualization purposes) and the featherec patch P; with invalid pixels marked blue (dark gray).
Figure 8: The error in the overlap region Sj, defined ir Equation 3 (left; the error-values S;(x) are normalized to the interval [0,1] for visualization purposes) and the featherec patch P; with invalid pixels marked blue (dark gray).
Figure 10: Traversal-map (Mj) construction by repeated morphological dilation of valid pixels. Level 1 shows valid pixels in gray, invalid pixels in black. Levels 2,...,n are the steps in creating Mj, each a zoom of the upper left corner of level 1. We then re-synthesize the pixels in each level of M; sequentially in ascending order. This is analogous to the method applied by an art restorer: a stepwise restoration of the hole from the boundary of the existing image.
Figure 10: Traversal-map (Mj) construction by repeated morphological dilation of valid pixels. Level 1 shows valid pixels in gray, invalid pixels in black. Levels 2,...,n are the steps in creating Mj, each a zoom of the upper left corner of level 1. We then re-synthesize the pixels in each level of M; sequentially in ascending order. This is analogous to the method applied by an art restorer: a stepwise restoration of the hole from the boundary of the existing image.
Figure 11: Re-synthesizing a single pixel (circle) in the over- lap region (with not yet synthesized pixels in white) using a 7x7 neighborhood. From the highlighted 7x7 area (left) we extract the image mask I (upper right) and the binary sup- port function J (lower right) for best pixel search in T. After Compositing P; into Rj, resulting in Rj composite (Fig. 1(b)), we re-synthesize invalid pixels in traversal-map ordering. In this procedure, each re-synthesized pixel is as- signed a square-shaped neighborhood (7x7 pixels in all of our examples) from which we gather already synthesized pixels into an image mask and a binary support function, identical to the best patch search of Section 2.1 (Fig. 11).
Figure 11: Re-synthesizing a single pixel (circle) in the over- lap region (with not yet synthesized pixels in white) using a 7x7 neighborhood. From the highlighted 7x7 area (left) we extract the image mask I (upper right) and the binary sup- port function J (lower right) for best pixel search in T. After Compositing P; into Rj, resulting in Rj composite (Fig. 1(b)), we re-synthesize invalid pixels in traversal-map ordering. In this procedure, each re-synthesized pixel is as- signed a square-shaped neighborhood (7x7 pixels in all of our examples) from which we gather already synthesized pixels into an image mask and a binary support function, identical to the best patch search of Section 2.1 (Fig. 11).
Figure 12: Pasting the final 3232 patch into a circularly shifted version of Figure 1(a). Top: MEBC "°, middle: feath- ering '7, bottom: overlap re-synthesis. See right image in Figure 5 for the above example before the overlap region is repaired.
Figure 12: Pasting the final 3232 patch into a circularly shifted version of Figure 1(a). Top: MEBC "°, middle: feath- ering '7, bottom: overlap re-synthesis. See right image in Figure 5 for the above example before the overlap region is repaired.
Figure 13: A comparison of texture synthesis algorithms on the benchmark green scales texture: (a) the input texture sample (64x64), (b) Efros/Leung synthesis, (c) Wei/Levoy synthesis, (d) Image Quilting, (e) Patch-Based Sampling, and (f) Hybrid Texture Synthesis. Each of the above is a 128 x 128 crop of a larger synthesis result. The synthesized textures behave toroidally in all cases, excluding the Image Quilting result. Patch sizes in (e) are 32X32. Parameters for (f): Snax = 0.02, Amax = 0.04, ov = 7, initial patch size 32 x 32.
Figure 13: A comparison of texture synthesis algorithms on the benchmark green scales texture: (a) the input texture sample (64x64), (b) Efros/Leung synthesis, (c) Wei/Levoy synthesis, (d) Image Quilting, (e) Patch-Based Sampling, and (f) Hybrid Texture Synthesis. Each of the above is a 128 x 128 crop of a larger synthesis result. The synthesized textures behave toroidally in all cases, excluding the Image Quilting result. Patch sizes in (e) are 32X32. Parameters for (f): Snax = 0.02, Amax = 0.04, ov = 7, initial patch size 32 x 32.
Figure 14: Comparing synthesis results: the original texture (left), Patch-Based Sampling (32 x32 patches, middle), and H) brid Texture Synthesis. The inputs are 200 x 200, results are all 256 x 256. In our results, 8max = 0.02, Amax = 0.02 ov = 12 and initial patch size 64x 64.
Figure 14: Comparing synthesis results: the original texture (left), Patch-Based Sampling (32 x32 patches, middle), and H) brid Texture Synthesis. The inputs are 200 x 200, results are all 256 x 256. In our results, 8max = 0.02, Amax = 0.02 ov = 12 and initial patch size 64x 64.
2 15: Hybrid Texture Synthesis results: the original texture (left), the tileable result (middle), and the result with the adaptive sampling grid overlaid (right). The inputs are 200x200 and 64x64, results are all 256X256, [8max,Amax, ov, initial patchsize| from top to bottom: [0.02,0.02, 14,64 x 64], [0.02,0.02, 14,64 x 64], [(0.01,0.05, 7, 32 x 32] and [0.04,0.02, 12,64 x 64]. The bottom row shows an example where our algorithm fails to produce a satisfying result.
2 15: Hybrid Texture Synthesis results: the original texture (left), the tileable result (middle), and the result with the adaptive sampling grid overlaid (right). The inputs are 200x200 and 64x64, results are all 256X256, [8max,Amax, ov, initial patchsize| from top to bottom: [0.02,0.02, 14,64 x 64], [0.02,0.02, 14,64 x 64], [(0.01,0.05, 7, 32 x 32] and [0.04,0.02, 12,64 x 64]. The bottom row shows an example where our algorithm fails to produce a satisfying result.

Related papers

Real-time texture synthesis by patch-based sampling
Harry Shum

ACM Transactions on Graphics, 2001

We present a patch-based sampling algorithm for synthesizing textures from an input sample texture. The patch-based sampling algorithm is fast. Using patches of the sample texture as building blocks for texture synthesis, this algorithm makes high-quality texture synthesis a real-time process. For generating textures of the same size and comparable (or better) quality, patch-based sampling is orders of magnitude faster than existing texture synthesis algorithms. The patch-based sampling algorithm synthesizes high-quality textures for a wide variety of textures ranging from regular to stochastic. By sampling patches according to a non-parametric estimation of the local conditional MRF density, we avoid mismatching features across patch boundaries. Moreover, the patch-based sampling algorithm remains effective when pixel-based non-parametric sampling algorithms fail to produce good results. For natural textures, the results of the patch-based sampling look subjectively better.

View PDFchevron_right
Multiscaled Texture Synthesis Using Multisized Pixel Neighborhoods
Xujiong Ye

IEEE Computer Graphics and Applications, 2000

View PDFchevron_right
Optimized tile-based texture synthesis
Jean-claude Paul

Proceedings of Graphics Interface 2007, 2007

One significant problem in tile-based texture synthesis is the presence of conspicuous seams in the tiles. The reason is that the sample patches employed as primary patterns of the tile set may not be well stitched if carelessly picked. In this paper, we introduce an optimized approach that can stably generate an ω-tile set of high pattern diversity and high quality. Firstly, an extendable rule is introduced to increase the number of sample patches to vary the patterns in an ω-tile set. Secondly, in contrast to the other concurrent techniques that randomly choose sample patches for tile construction, our technique uses Genetic Algorithm to select the feasible patches from the input example. This operation insures the quality of the whole tile set. Experimental results verify the high quality and efficiency of the proposed algorithm.

View PDFchevron_right
Patchmatch for Texture Synthesis
Stefan Jeschke

2011

This paper is a report about a project practical course (“Projektpraktikum”) done at the Institute for Computer Graphics and Algorithms of the Vienna University of Technology. Advisor: Stefan Jeschke. The goal of this course was to implement a texture synthesis algorithm and analyze its performance and visual quality. Two papers form the theoretical foundation of this project: “Texture Optimization for Example-based Synthesis” by Kwatra et al [KEBK05] and “Patchmatch: A randomized Correspondence Algorithm for Structural Image Editing” by Barnes et al [BSFG09].

View PDFchevron_right
Multiscale texture synthesis
Eitan Grinspun

ACM Transactions on Graphics, 2008

Example-based texture synthesis algorithms have gained widespread popularity for their ability to take a single input image and create a perceptually similar non-periodic texture. However, previous methods rely on single input exemplars that can capture only a limited band of spatial scales. For example, synthesizing a continent-like appearance at a variety of zoom levels would require an impractically high input resolution. In this paper, we develop a multiscale texture synthesis algorithm. We propose a novel example-based representation, which we call an exemplar graph, that simply requires a few low-resolution input exemplars at different scales. Moreover, by allowing loops in the graph, we can create infinite zooms and infinitely detailed textures that are impossible with current example-based methods. We also introduce a technique that ameliorates inconsistencies in the user&#39;s input, and show that the application of this method yields improved interscale coherence and highe...

View PDFchevron_right
A fast texture synthesis technique using spatial neighborhood
Muath Sabha

2005

In this technical report, we present a texture synthesis technique that stores all possible histories in memory structure acting as a Finite State Machine. The histories, which are the most probable spatial neighborhood for each pixel or patch, are stored in each state of the FSM. After constructing that structure, the synthesis goes very fast and smooth in a raster scan order. We found that our technique works well for a wide variety of textures, ranging from stochastic to regular structured textures, including near-regular textures. We show that our technique is faster, and gives better results than previous pixel- or patch-based techniques, with minimal user input.

View PDFchevron_right
A Survey of Texture Synthesis Various Techniques
IJSRD - International Journal for Scientific Research and Development

IJSRD, 2014

Texture synthesis is a latest technology which grows rapidly in recent ten years. Texture synthesis is a main process of computer graphics applications, which can enhance the realistic rendering greatly. Single texture synthesis cannot meet the customer need because it is straightforward approach. Multi-exemplars synthesis is a challenging research topic to increase the richness of texture details. We survey the algorithms for synthesizing textures that have proposed over the synthesized textures are given. The goal is to further understand textures, to provide the methodology needed to generate various textures in other studies, and pursuing better models for synthesizing more interesting textures.

View PDFchevron_right
Refining structural texture synthesis approach
Mukul Sutaone

Structural textures are characterised by a repeating pattern called 'Texton' and placement rule -that determines the nature of periodicity. Based on the periodicity, textures are classified as homogeneous -perfectly periodic -and weakly homogeneous -quasi-periodic. Both of these textures are assumed to be combination of structural information, illumination, that is, average brightness at different sites of the texture and stochasticity to allow local variations. A top-down approach extracts structural information, that is, the grid, representative texton and illumination component from the original texture patch and then the information is used to synthesise similar textures. Obviously, the technique does not produce ditto copies in the synthesised texture. The experimentation is carried out to improve quality of synthesised textures by incorporating multiple representative textons. Also a parameter, namely homogeneity co-efficient (HC), is suggested to compare the original texture patch and synthesised texture. The parameter captures variations in the textons, contents and sizes both, and thus can be used to compare the synthesis results. The suitability of the proposed synthesis approach and HC is verified by rigorous experimentation on weakly homogeneous artificial and standard structural textures. Efforts are also made to utilise the multicore processing capability of the processor to improve the speed of analysis and synthesis phases.

View PDFchevron_right
Image replacement through texture synthesis
Lucas Pereira

1997

Photographs and images often have regions which are in some sense flawed. Often, there may be a stain or an undesired feature covering a significant portion of the image, and an algorithm that can "fix" the image is desired. We propose a technique based on Heeger and Bergen's texture synthesis algorithm . By integrating a composition step into the aforementioned algorithm, one portion of the image can be substituted with a synthetic texture derived from another portion of the image. The goal of this synthetic texture is to create a plausible patch without visible seams or repetitive features. We compare this technique, which works on areas that are "stochastic" in nature, with other variations of image replacement which utilize a combination of image composition and texture synthesis.

View PDFchevron_right
A Roller - Fast Sampling-Based Texture Synthesis Algorithm
Michal Haindl

This paper describes a method for synthesizing natural textures that realistically matches given colour texture ap- pearance. The novel texture synthesis method, which we call the roller, is based on the overlapping tiling and subsequent minimum error boundary cut. An optimal double toroidal patch is seamlessly repeated during the syn- thesis step. While the method allows only moderate texture compression it is extremely fast due to separation of the analytical step of the algorithm from the texture synthesis part, universal, and easily implementable in a graph- ical hardware for purpose of real-time colour texture rendering.

View PDFchevron_right

References (33)

  1. M. Ashikhmin. Synthesizing Natural Textures. Proc. ACM Symposium on Interac- tive 3D Graphics, 217-226, 2001. 2, 7
  2. Z. Bar-Joseph, R. El-Yaniv, D. Lischinski and M. Werman. Texture Mixing and Texture Movie Synthesis Using Statistical Learning. IEEE Transactions on Visual- ization and Computer Graphics, 120-135, 2001. 2
  3. J. Blinn and M. Newell. Texture and reflection in computer generated images. Communications of the ACM, 19(10): 542-547, October 1976. 1
  4. J.S. De Bonet. Multiresolution Sampling Procedure for Analysis and Synthesis of Texture Images. SIGGRAPH 97 Conference Proceedings, 361-368, 1997. 2
  5. O. Cuisenaire. Region growing Euclidean distance transforms. ICIAP'97, 9th In- ternational Conference on Image Analysis and Processing, , 1997. 6
  6. J.-M. Dischler and D. Ghazanfarpour. A Geometrical Based Method for Highly Complex Structured Textures Generation. Computer Graphics Forum, 14(4): 203- 215, October 1995. 3
  7. J.-M. Dischler, K. Maritaud, B. Lévy and D. Ghazanfarpour. Texture Particles. EUROGRAPHICS 2002 Conference Proceedings, 2002. 3
  8. I. Drori, D. Cohen-Or and H. Yeshurun. Example-Based Style Synthesis. to appear in IEEE Computer Society Conference on Computer Vision and Pattern Recogni- tion, 2003. 3
  9. A. Efros and T. Leung. Texture Synthesis by Non-parametric Sampling. IEEE International Conference on Computer Vision, 1033-1038, 1999. 2 c The Eurographics Association 2003.
  10. A.A. Efros and W.T. Freeman. Image Quilting for texture synthesis and transfer. SIGGRAPH 2001 Conference Proceedings, 341-346, 2001. 1, 2, 3, 4, 7
  11. J.D. Foley, A. van Dam, S.K. Feiner and J.F. Hughes Computer Graphics. Princi- ples and Practice. Addison Wesley, 1990. 4
  12. B. Guo, H. Shum and Y.-Q. Xu. Chaos Mosaic: Fast and Memory Efficient Texture Synthesis. Microsoft Technical Report MSR-TR-2000-32, 2001. 2
  13. D.J. Heeger and J.R. Bergen. Pyramid-Based Texture Analysis/Synthesis. SIG- GRAPH 95 Conference Proceedings, 229-238, 1995. 2
  14. A. Hertzmann, C.E. Jacobs, N. Oliver, B. Curless and D.H. Salesin. Image Analo- gies. ACM Computer Graphics (Proc. of SIGGRAPH '01), 327-340, 2001. 2
  15. T. Leung and J. Malik. Recognizing Surfaces using Three-Dimensional Textons. Proc. International Conference in Computer Vision, 1010-1017, 1999. 3
  16. T. Leung and J. Malik. Representing and recognizing the visual appearance of materials using 3d textons. International Journal of Computer Vision, 43(1): 29- 44, June 2001. 3
  17. L. Liang. C. Liu, Y.-Q. Xu, B. Guo and H.-Y. Shum. Real-time Texture Synthesis by Patch-Based Sampling. ACM Transactions on Graphics, 20(3): 127-150, July 2001. 1, 2, 3, 4, 5, 6, 7
  18. Y. Liu and Y. Tsin. The Promise and Perils of Near-Regular Texture. Texture 2002, 2002. 8
  19. X. Liu, Y. Yu and H.-Y. Shum. Synthesizing Bidirectional Texture Functions for Real-World Surfaces. SIGGRAPH 2001, Computer Graphics Proceedings, 97-106, 2001. 2, 8
  20. K. Perlin. An Image Synthesizer. SIGGRAPH 1985, Computer Graphics Proceed- ings, 287-296, 1985. 1
  21. E. Praun, A. Finkelstein and H. Hoppe. Lapped Textures. Siggraph 2000, Computer Graphics Proceedings, 465-470, 2000. 2
  22. J. Portilla and E.P. Simoncelli. A Parametric Texture Model based on Joint Statistics of Complex Wavelet Coefficients. Int'l Journal of Computer Vision, 40(1): 49-71, 2000. 2
  23. C. Soler, M.-P. Cani and A. Angelidis. Hierarchical Pattern Mapping. SIGGRAPH 2002 Conference Proceedings, 673-680, 2002. 1, 2, 3, 4
  24. X. Tong, J. Zhang, L. Liu, X. Wang, B. Guo and H.-Y. Shum. Synthesis of Bidi- rectional Texture Functions on Arbitrary Surfaces. SIGGRAPH 2002 Conference Proceedings, 665-672, 2002. 2, 7
  25. G. Turk. Texture Synthesis on Surfaces. SIGGRAPH 2001, Computer Graphics Proceedings, 347-354, 2001. 2
  26. L.-Y. Wei and M. Levoy. Fast Texture Synthesis Using Tree-Structured Vector Quantization. Siggraph 2000, Computer Graphics Proceedings, 479-488, 2000. 2, 7
  27. L.-Y. Wei and M. Levoy. Texture Synthesis Over Arbitrary Manifold Surfaces. SIGGRAPH 2001, Computer Graphics Proceedings, 355-360, 2001. 2
  28. L. Ying, A. Hertzmann, H. Biermann and D. Zorin. Texture and Shape Synthesis on Surfaces. Proc. 12th Eurographics Workshop on Rendering, 301-312, 2001. 2
  29. S. Zelinka and M. Garland. Towards Real-Time Texture Synthesis with the Jump Map. Proc. 13th Eurographics Workshop on Rendering, 2002. 2
  30. S.C. Zhu, Y. Wu and D. Mumford. Minimax Entropy Principles and Its Application to Texture Modeling. NeurComp, 9(8): 1627-1660, 1997. 3
  31. S.C. Zhu, Y. Wu and D. Mumford. FRAME: Towards a Unified Theory for Texture Modeling. IJCV, 27(2): 107-126, March 1998. 3
  32. S.C. Zhu, C.E. Guo, Y. Wu and Y. Wang. What Are Textons?. ECCV02, 793 ff., 2002. 3
  33. c The Eurographics Association 2003.

Related papers

Fast and High Quality Overlap Repair for Patch-Based Texture Synthesis
Andy Nealen

Computer Graphics International, 2004. Proceedings, 2004

View PDFchevron_right
Sample-Based Texture Synthesis
Muath Sabha

The goal of this dissertation is to attain a fast texture synthesis technique able to generate visually pleasing results from a sample texture for a wide range of texture categories. Computer graphics applications often use textures to decorate virtual objects in order to increase realism without modifying geometric details. In recent years, texture synthesis has been a useful tool to generate large textures starting from a sample texture. Texture synthesis is also used to fill holes in damaged pictures, or to generate illimitable textures in interactive applications such as games and simulators. In our dissertation, we present our contributions in texture synthesis and editing. The contributions are in three directions, a pixel-based texture synthesis technique, a patch-based texture synthesis technique, and a patch-based texture synthesis and editing technique. For the Pixel-Based technique, we were inspired by the observation that the most common artefacts which occur when synthe...

View PDFchevron_right
Volume xx (200y), Number z, pp. 111 Texture Synthesis using Exact Neighborhood Matching
Muath Sabha

2016

In this paper we present an elegant pixel-based texture synthesis technique which is able to generate visually pleasing results from source textures of both stochastic and structured nature. Inspired by the observation that the most common artefacts which occur when synthesizing textures are high frequency discontinuities, our technique tries to avoid these artefacts by forcing at least one of the direct neighboring pixels in each causal neighborhood to match within a predetermined threshold. This does not only avoid deterioration of the visual quality, but also results in faster synthesis timings. We demonstrate our technique on a variety of stochastic and structured textures.

View PDFchevron_right
Local Patch Based Texture Synthesis
Anish Kumar

www.ciitresearch.org/dip.html, 2012

Texture synthesis is growing a sample texture to an infinite texture which looks visually similar to the sample texture. Texture synthesis is needed to seamlessly cover areas and surfaces with a texture. A large number of patch based texture synthesis algorithm have been developed. They provide better result and the speed of the algorithm is quite high. This project proposes a patch based texture synthesis approach along a spiral path. This algorithm searches for new matching patch among the neighborhood of the previous matched patch along a spiral path and this process is continued till the image is fully filled.Since the search space is reduced ,a synthesis speed-up is obtained. This approach provides better quality and high speed of execution. Experiments result shows that this algorithm gives better quality.

View PDFchevron_right
Towards effective evaluation of geometric texture synthesis algorithms
Zainab AlMeraj

Proceedings of the Symposium on Non-Photorealistic Animation and Rendering - NPAR '13, 2013

In recent years, an increasing number of example-based Geometric Texture Synthesis (GTS) algorithms have been proposed. However, there have been few attempts to evaluate these algorithms rigorously. We are driven by this lack of validation and the simplicity of the GTS problem to look closer at perceptual similarity between geometric arrangements. Using samples from a geological database, our research first establishes a dataset of geometric arrangements gathered from multiple synthesis sources. We then employ the dataset in two evaluation studies. Collectively these empirical methods provide formal foundations for perceptual studies in GTS, insight into the robustness of GTS algorithms and a better understanding of similarity in the context of geometric texture arrangements.

View PDFchevron_right

Related topics

  • Texture Synthesis
  • Hybrid Algorithm
  • Morphological Operation
    • 580 California St., Suite 400
    San Francisco, CA, 94104
    © 2025 Academia. All rights reserved