Academia.eduAcademia.edu

Outline

Title
Abstract
Figures
Introduction
Results and Discussions
References
All Topics
Computer Science
Computer Graphics

Footprint area sampled texturing

Profile image of Frank DachilleFrank Dachille

2004, IEEE Transactions on Visualization and Computer Graphics

Abstract

We study texture projection based on a four region subdivision: magnification, minification, and two mixed regions. We propose improved versions of existing techniques by providing exact filtering methods which reduce both aliasing and overblurring, especially in the mixed regions. We further present a novel texture mapping algorithm called FAST (Footprint Area Sampled Texturing), which not only delivers high quality, but also is efficient. By utilizing coherence between neighboring pixels, performing prefiltering, and applying an area sampling scheme, we guarantee a minimum number of samples sufficient for effective antialiasing. Unlike existing methods (e.g., MIP-map, Feline), our method adapts the sampling rate in each chosen MIP-map level separately to avoid undersampling in the lower level l for effective antialiasing and to avoid oversampling in the higher level l þ 1 for efficiency. Our method has been shown to deliver superior image quality to Feline and other methods while retaining the same efficiency. We also provide implementation trade offs to apply a variable degree of accuracy versus speed.

Figures (17)
Fig. 1. Four regions of a texture mapped checkerboard plane magnification (cyan), minification (red), and mixed regions (green anc blue).
Fig. 1. Four regions of a texture mapped checkerboard plane magnification (cyan), minification (red), and mixed regions (green anc blue).
Fig. 3. No texel may be projected into a pixel footprint in naive forwar point projection methods.
Fig. 3. No texel may be projected into a pixel footprint in naive forwar point projection methods.
Fig. 2. Forward projected texel footprint may cover no pixels in naive forward footprint projection methods.
Fig. 2. Forward projected texel footprint may cover no pixels in naive forward footprint projection methods.
Fig. 4. Forward point projection texture filtering with (a) backward bilinear interpolation to fill in the holes [10] (notice aliasing due to subsampling in one direction) and (b) our solution with supersampling in texture space to fill in the holes. (c) and (d) are zoom-ins of the marked rectangles of (a) and (b), respectively.
Fig. 4. Forward point projection texture filtering with (a) backward bilinear interpolation to fill in the holes [10] (notice aliasing due to subsampling in one direction) and (b) our solution with supersampling in texture space to fill in the holes. (c) and (d) are zoom-ins of the marked rectangles of (a) and (b), respectively.
Fig. 5. The backward projected footprint may cover no texels in naive backward footprint projection methods.
Fig. 5. The backward projected footprint may cover no texels in naive backward footprint projection methods.
Fig. 6. Solutions to the undersampling problem in backward footprint projection: (a) overblurred EWA [15] and (b) out clamping method. (c) and (d) are zoom-ins of the marked rectangles of (a) and (b), respectively.
Fig. 6. Solutions to the undersampling problem in backward footprint projection: (a) overblurred EWA [15] and (b) out clamping method. (c) and (d) are zoom-ins of the marked rectangles of (a) and (b), respectively.
Fig. 7. The minor axis is over supersampled in naive backward poin projection methods.
Fig. 7. The minor axis is over supersampled in naive backward poin projection methods.
Fig. 8. Prefiltering methods: (a) MIP-mapping; (b) Footprint Assembly/ Feline.
Fig. 8. Prefiltering methods: (a) MIP-mapping; (b) Footprint Assembly/ Feline.
Fig. 9. Comparison between Footprint Assembly and Feline: (a) screen space, (b) texture space of Footprint Assembly, (c) and Feline
Fig. 9. Comparison between Footprint Assembly and Feline: (a) screen space, (b) texture space of Footprint Assembly, (c) and Feline
To hold the partial contribution of all subpixels during the scanline processing, intuitively, two scanline buffers are required: one for the current scanline L; and another for the next scanline L;,,. But, for a compact implementation, a scanline buffer of size length +1 working as a FIFO is sufficient, where length is the length of the scanline. Once the final colors are produced for pixels in scanline L;, their memory can be immediately released for storing the partial colors of pixels in scanline L;.;. Each pixel of the scanline buffer stores the accumulated color RGB and weight. The (u,v) coordinates in texture space are also stored in the buffer. This is used for pixel vector calculation, to be discussed in the next section. Since the operations are conducted in two MIP-map levels, hence, two sets of buffers are needed, resulting in two scanline buffers. This is still feasible for hardware implementation if we use bucket When the pixel coverage overlaps, subpixels between neighboring pixels contribute to all surrounding pixels. For example, in Fig. 11a, subpixels within the square region aefg contribute to pixels a, e, f, and g. The naive implementation of our adaptive sampling method (Section 2.4) calculates the subpixels of each pixel sepa- rately, which means that the subpixels are calculated four times. This requires not only costly computation, but also a high memory bandwidth. To address this issue, coherence
To hold the partial contribution of all subpixels during the scanline processing, intuitively, two scanline buffers are required: one for the current scanline L; and another for the next scanline L;,,. But, for a compact implementation, a scanline buffer of size length +1 working as a FIFO is sufficient, where length is the length of the scanline. Once the final colors are produced for pixels in scanline L;, their memory can be immediately released for storing the partial colors of pixels in scanline L;.;. Each pixel of the scanline buffer stores the accumulated color RGB and weight. The (u,v) coordinates in texture space are also stored in the buffer. This is used for pixel vector calculation, to be discussed in the next section. Since the operations are conducted in two MIP-map levels, hence, two sets of buffers are needed, resulting in two scanline buffers. This is still feasible for hardware implementation if we use bucket When the pixel coverage overlaps, subpixels between neighboring pixels contribute to all surrounding pixels. For example, in Fig. 11a, subpixels within the square region aefg contribute to pixels a, e, f, and g. The naive implementation of our adaptive sampling method (Section 2.4) calculates the subpixels of each pixel sepa- rately, which means that the subpixels are calculated four times. This requires not only costly computation, but also a high memory bandwidth. To address this issue, coherence
rendering [19], in which the framebuffer is subdivided into coherent regions that are rendered independently (also known as tiled rendering or chunking). Bucket rendering enables a deeper working set memory to be kept on chip.
rendering [19], in which the framebuffer is subdivided into coherent regions that are rendered independently (also known as tiled rendering or chunking). Bucket rendering enables a deeper working set memory to be kept on chip.
The magnitude of the minor vector m can be approxi- mated by: of an additonal pixel to its right. However, note that, for these additional pixels, only their projections are calculated. The magnitudes of these vectors are approximated with: The minor vector m is used to compute the MIP level / and the scaling factor f of level I:
The magnitude of the minor vector m can be approxi- mated by: of an additonal pixel to its right. However, note that, for these additional pixels, only their projections are calculated. The magnitudes of these vectors are approximated with: The minor vector m is used to compute the MIP level / and the scaling factor f of level I:
Fig. 13. (a) Feline versus (b) our FAST area sampling. FAST area sampling method, the minor axis rz is used to determine the MIP-map level (5), which is 1 for this case. In level 1, the scaled two vectors become |r{| = 8.0 and \r5| = 1.2, respectively. The total sampling rate is deter- mined as N = [8.0 x 1.2] = 10. The number of samples in level 2 is N’ = [#7] = 3. In total, Feline spends 14 samples for the two MIP-map levels, while FAST spends 13 samples. Straightforward implementation of the adaptive sampling method creates the highest quality image with less blurring and better antialiasing (Fig. 14b). Utilizing coherence only in the adaptive sampling method (without prefiltering) promises the same high quality (Fig. 14c). Feline delivers better quality than MIP-mapping by introducing anisotro- pic filtering (Fig. 14d). This can be observed from the text image, but aliasing is still present in the checkerboard image. By setting the clamping factor to 4, the quality is similar (Fig. 14e). However, by further reducing the clamping factor to 2, the image becomes blurrier (Fig. 14f). Our FAST method, with coherence utilization, prefiltering, and the footprint area sampling scheme, creates images of indistinguishable quality as compared to straightforward adaptive sampling (Fig. 14g). By setting the clamping factor to 4, the quality is similar (Fig. 14h). However, by setting the clamping factor to 2, the image becomes a bit blurrier and more aliased (Fig. 14i). The difference can be seen more clearly from zoom-ins in Figs. 14j-l. The quality of animations is consistent with that of the extracted frames for each method.
Fig. 13. (a) Feline versus (b) our FAST area sampling. FAST area sampling method, the minor axis rz is used to determine the MIP-map level (5), which is 1 for this case. In level 1, the scaled two vectors become |r{| = 8.0 and \r5| = 1.2, respectively. The total sampling rate is deter- mined as N = [8.0 x 1.2] = 10. The number of samples in level 2 is N’ = [#7] = 3. In total, Feline spends 14 samples for the two MIP-map levels, while FAST spends 13 samples. Straightforward implementation of the adaptive sampling method creates the highest quality image with less blurring and better antialiasing (Fig. 14b). Utilizing coherence only in the adaptive sampling method (without prefiltering) promises the same high quality (Fig. 14c). Feline delivers better quality than MIP-mapping by introducing anisotro- pic filtering (Fig. 14d). This can be observed from the text image, but aliasing is still present in the checkerboard image. By setting the clamping factor to 4, the quality is similar (Fig. 14e). However, by further reducing the clamping factor to 2, the image becomes blurrier (Fig. 14f). Our FAST method, with coherence utilization, prefiltering, and the footprint area sampling scheme, creates images of indistinguishable quality as compared to straightforward adaptive sampling (Fig. 14g). By setting the clamping factor to 4, the quality is similar (Fig. 14h). However, by setting the clamping factor to 2, the image becomes a bit blurrier and more aliased (Fig. 14i). The difference can be seen more clearly from zoom-ins in Figs. 14j-l. The quality of animations is consistent with that of the extracted frames for each method.
Fig. 14. Texture mapping using (a) traditional MIP-map, (b) adaptive sampling without coherence, (c) adaptive sampling with coherence, (d) Feline without clamping, (e) Feling with clamp = 4, (f) Feline with clamp = 2, (g) FAST without clamping, (h) FAST with clamp = 4, (i) FAST with clamp = 2. (j), (kK), and (I) are zoom-ins of the marked racteangles of (a), (d) and (g), respectively.
Fig. 14. Texture mapping using (a) traditional MIP-map, (b) adaptive sampling without coherence, (c) adaptive sampling with coherence, (d) Feline without clamping, (e) Feling with clamp = 4, (f) Feline with clamp = 2, (g) FAST without clamping, (h) FAST with clamp = 4, (i) FAST with clamp = 2. (j), (kK), and (I) are zoom-ins of the marked racteangles of (a), (d) and (g), respectively.
Fig. 15. Average rendering time per pixel (milliseconds) for 110 frames of a room scene animation.
Fig. 15. Average rendering time per pixel (milliseconds) for 110 frames of a room scene animation.
Fig. 16. Number of sample points per pixel for 110 frames of a room scene animation.
Fig. 16. Number of sample points per pixel for 110 frames of a room scene animation.

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
TextureMontage: Seamless Texturing of Arbitrary Surfaces From Multiple Images
kun zhou

2005

We propose a technique, called TextureMontage, to seamlessly map a patchwork of texture images onto an arbitrary 3D model. A texture atlas can be created through the specification of a set of correspondences between the model and any number of texture images. First, our technique automatically partitions the mesh and the images, driven solely by the choice of feature correspondences. Most charts will then be parameterized over their corresponding image planes through the minimization of a distortion metric based on both geometric distortion and texture mismatch across patch boundaries and images. Lastly, a surface texture inpainting technique is used to fill in the remaining charts of the surface with no corresponding texture patches. The resulting texture mapping satisfies the (sparse or dense) user-specified constraints while minimizing the distortion of the texture images and ensuring a smooth transition across the boundaries of different mesh patches.

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
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
Feature-aware texturing
Ran Gal

2006

We present a method for inhomogeneous 2D texture mapping guided by a feature mask, that preserves some regions of the image, such as foreground objects or other prominent parts. The method is able to arbitrarily warp a given image while preserving the shape of its features by constraining their deformation to be a similarity transformation. In particular, our method allows global or local changes to the aspect ratio of the texture without causing undesirable shearing to the features. The algorithmic core of our method is a particular formulation of the Laplacian editing technique, suited to accommodate similarity constraints on parts of the domain. The method is useful in digital imaging, texture design and any other applications involving image warping, where parts of the image have high familiarity and should retain their shape after modification.

View PDFchevron_right
Uniform texture synthesis and texture mapping using global parameterization
Shing Toung Yau

The Visual Computer, 2005

Texture mapping and texture synthesis are two popular methods for the decoration of surfaces with visual detail. Here, an existing challenge is to preserve, or at least balance, two competing metrics: scale and angle. In this paper we present two methods for this, both based on global conformal parameterization. First, we describe a texture synthesis algorithm for surfaces with arbitrary topology. By using the conformal parameterization, the 3D surface texture synthesis problem can be converted to a 2D image synthesis problem, which is more intuitive, easier, and conceptually simpler. While the conformality of the parameterization naturally preserves the angles of the texture, in this paper we provide a multi-scale technique to also maintain a more uniform area scaling factor. A second novel contribution is to employ the global parameterization to simultaneously preserve orthogonality and size in texture mapping. For this, we show that a conformal factor-driven mass-spring method offers a convenient way to trade off these two qualitative metrics. Our algorithms are simple, efficient and automatic, and they are theoretically sound and universal to general surfaces as well.

View PDFchevron_right
Approximation techniques for high performance texture mapping
Mehmet Demirer

Computers & Graphics, 1996

Accurate perspective mapping in real-time requires costly division operations per pixel and therefore ap proximation techniques are often employed. These per mit the mapping to be performed by interpolation, but generally with a significant set-up cost for the computa tion of the parameters. An efficient approximation tech nique which achieves good results with modest set-up requirements is presented. The technique uses Cheby shev control points to minimise errors.

View PDFchevron_right
Antialiased parameterized solid texturing simplified for consumer-level hardware implementation. 1999 SIGGRAPH
Terrance Coleman

Eurographics Workshop on …, 1999

Procedural solid texturing was introduced fourteen years ago, but has yet to find its way into consumer level graphics hardware for real-time operation. To this end, a new model is introduced that yields a parameterized function capable of synthesizing the most common procedural solid textures, specifically wood, marble, clouds and fire. This model is simple enough to be implemented in hardware, and can be realized in VLSI with as little as 100,000 gates. The new model also yields a new method for antialiasing synthesized textures. An expression for the necessary box filter width is derived as a function of the texturing parameters, the texture coordinates and the rasterization variables. Given this filter width, a technique for efficiently box filtering the synthesized texture by either mip mapping the color table or using a summed area color table are presented. Examples of the antialiased results are shown.

View PDFchevron_right
Real Time Image Enhancement Using Texture Synthesis (RETS)
Sunil Pathak

—Real time Enhancement using texture synthesis combines interpolation, classification and patch based texture synthesis to enhance low resolution imagery. RETS uses a low resolution source image as input and several high resolution sample textures. The output of RETS is a high resolution image with the structure of the source image, but with detail consistent with the high resolution sample textures. Image Interpolation: Interpolation is the primary technique used for image scaling. Image scaling is the process of taking a source image and extending it to create a large image. The primary problem with enlarging images using interpolation is that the large result contains the same amount of discrete data as smaller source image [25]. Two types of interpolation are bilinear and bicubic. Bilinear interpolation uses 2x2 neighbourhoods of data points to calculate pixel color between data points. Bicubic interpolation uses 4x4 neighbourhoods of data points to calculate pixel color between data points [16].Using texture Synthesis solve two problems • It allows to user to specify the detail to be inserted into the output image by providing a representative sample for the system to replicate. • Applying texture synthesis appropriately will allow us to avoid the unnatural repetition of texture tiles that can occur with standard texture mapping.

View PDFchevron_right
Fast multi-scale joint bilateral texture upsampling
Wei Hua

The Visual Computer, 2010

We present a new approach using a multi-scale joint bilateral filter for upsampling the synthesized texture generated by optimization-based methods. Our method is based on the following motivation: if the available exemplar texture is used as a priority to upsample the synthesized texture, a high resolution result that prevents image blurring can be obtained. Our multi-scale joint bilateral upsampling applies a spatial filter on each multi-scale layer of the synthesized texture, and jointly applies a similar range filter on exemplar texture, which guides the interpolation from low to high resolution, by magnifying and combining the upsampled information; the details of the upsampled texture are progressively enhanced, and the image blurring artifacts can be effectively avoided. We offer an accelerated joint bilateral filter, which enables our upsampling method to interactively generate a large texture. In addition, we propose a detail-aware texture optimization approach that incorporates image detail in texture optimization to improve the quality of the synthesized texture, on which the multi-scale joint bilateral filter works to generate a more convincing result. We show results for upsampling image and video textures and compare them to traditional upsampling methods, by this demonstrating that with low computational and memory costs, our method achieves better results.

View PDFchevron_right

References (23)

  1. A.C. Barkans, "High-Quality Rendering Using the Talisman Architecture," Proc. 1997 SIGGRAPH/Eurographics Workshop Graphics Hardware, pp. 79-88, Aug. 1997.
  2. J.F. Blinn and M.E. Newell, "Texture and Reflection in Computer Generated Images," Comm. ACM, vol. 19, pp. 542-546, 1976.
  3. D. Blythe and T. McReynolds, "Advanced Graphics Programming Techniques Using OpenGL," SIGGRAPH '99 Course, 1999.
  4. B. Chen, F. Dachille, and A. Kaufman, "Forward Image Warping," Proc. IEEE Visualization '99, pp. 89-96, Oct. 1999.
  5. F.C. Crow, "Summed-Area Tables for Texture Mapping," Compu- ter Graphics (SIGGRAPH '84 Proc.), H. Christiansen, ed., vol. 18, pp. 207-212, July 1984.
  6. J.P. Ewins, M.D. Waller, M. White, and P.F. Lister, "Mip-Map Level Selection for Texture Mapping," IEEE Trans. Visualization and Computer Graphics, vol. 4, no. 4, pp. 317-329, Oct.-Dec. 1998.
  7. E.A. Feibush, M. Levoy, and R.L. Cook, "Synthetic Texturing Using Digital Filters," Computer Graphics (SIGGRAPH '80 Proc.), vol. 14, no. 3, pp. 294-301, July 1980.
  8. A. Fournier and E. Fiume, "Constant-Time Filtering with Space- Variant Kernels," Computer Graphics (SIGGRAPH '88 Proc.), vol. 22, no. 4, pp. 229-238, Aug. 1988.
  9. M. Gangnet, D. Perny, and P. Coueignoux, "Perspective Mapping of Planar Textures," Computers and Graphics, vol. 8, no. 2, pp. 115- 123, 1984.
  10. D. Ghazanfarpour and B. Peroche, "A High-Quality Filtering Using Forward Texture Mapping," Computers and Graphics, vol. 15, no. 4, pp. 569-577, 1991.
  11. A. Glassner, "Adaptive Precision in Texture Mapping," Computer Graphics (SIGGRAPH '86 Proc.), vol. 20, pp. 297-306, Aug. 1986.
  12. N. Greene and P.S. Heckbert, "Creating Raster Omnimax Images from Multiple Perspective Views Using the Elliptical Weighted Average Filter," IEEE Computer Graphics and Applications, vol. 6, no. 6, pp. 21-27, June 1986.
  13. P. Haeberli and K. Akeley, "The Accumulation Buffer: Hardware Support for High-Quality Rendering," Computer Graphics (SIG- GRAPH '90 Proc.), vol. 24, no. 4, pp. 309-318, Aug. 1990.
  14. P.S. Heckbert, "Survey of Texture Mapping," IEEE Computer Graphics and Applications, vol. 6, no. 11, pp. 56-67, Nov. 1986.
  15. P.S. Heckbert, "Fundamentals of Texture Mapping and Image Warping," MS thesis, Dept. of Electrical Eng. and Computer Science, Univ. of California, Berkeley, June 1989.
  16. J. McCormack, R. Perry, K.I. Farkas, and N.P. Jouppi, "Feline: Fast Elliptical Lines for Anisotropic Texture Mapping," Computer Graphics (SIGGRAPH '99 Proc.), pp. 243-250, Aug. 1999.
  17. A. Schilling, G. Knittel, and W. Strasser, "Texram: Smart Memory for Texturing," IEEE Computer Graphics and Applications, vol. 16, no. 3, pp. 32-41, May 1996.
  18. J.E. Swan II, K. Mueller, T. Mo ¨ller, N. Shareef, R.A. Crawfis, and R. Yagel, "An Anti-Aliasing Technique for Splatting," Proc. IEEE Visualization '97, pp. 197-204, Nov. 1997.
  19. J. Torborg and J. Kajiya, "Talisman: Commodity Real-Time 3D Graphics for the PC," Computer Graphics (SIGGRAPH '96), pp. 353- 364, Aug. 1996.
  20. L. Williams, "Pyramidal Parametrics," Computer Graphics (SIG- GRAPH '83 Proc.), vol. 17, no. 3, pp. 1-11, July 1983.
  21. S. Winner, M. Kelley, B. Pease, B. Rivard, and A. Yen, "Hardware Accelerated Rendering of Antialiasing Using a Modified A-Buffer Algorithm," Computer Graphics (SIGGRAPH '97 Proc.), vol. 31, no. 3A, pp. 307-316, Aug. 1997.
  22. G. Wolberg, Digital Image Warping. Los Alamitos, Calif.: IEEE CS Press, 1990.
  23. M. Zwicker, H. Pfister, J. van Baar, and M. Gross, "EWA Volume Rendering," Proc. IEEE Visualization '01, pp. 29-36, Oct. 2001.

Related papers

Space-Optimized Texture Maps
Laurent Balmelli

Computer Graphics Forum, 2002

Texture mapping is a common operation to increase the realism of three-dimensional meshes at low cost. We propose a new texture optimization algorithm based on the reduction of the physical space allotted to the texture image. Our algorithm optimizes the use of texture space by computing a warping function for the image and new texture coordinates. Neither the mesh geometry nor its connectivity are modified by the optimization. Our method uniformly distributes frequency content of the image in the spatial domain. In other words, the image is stretched in high frequency areas, whereas low frequency regions are shrunk. We also take into account distortions introduced by the mapping onto the model geometry in this process. The resulting image can be resampled at lower rate while preserving its original details. The unwarping is performed by the texture mapping function. Hence, the spaceoptimized texture is stored as-is in texture memory and is fully supported by current graphics hardware. We present several examples showing that our method significantly decreases texture memory usage without noticeable loss in visual quality.

View PDFchevron_right
MIP-Map Level Selection for Texture Mapping
Paul Lister

IEEE Transactions on Visualization and Computer Graphics, 1998

Texture mapping is a fundamental feature of computer graphics image generation. In current PC based acceleration hardware, MIP-mapping with bilinear and trilinear filtering are commonly used filtering techniques for reducing spatial aliasing artifacts. The effectiveness of these techniques in reducing image aliasing at the expense of blurring is dependent upon the MIP-map level selection and the associated calculation of screen space to texture space pixel scaling. This paper describes an investigation of practical methods for per pixel and per primitive level of detail calculation. This investigation was carried out as part of the design work for a screen space rasterization ASIC. The implementations of several algorithms of comparable visual quality are discussed and a comparison is provided in terms of per primitive and per pixel computational costs.

View PDFchevron_right
Space-Optimized Texture Maps (Guenter Enderle [Best Paper] Award
Laurent Balmelli

Computer Graphics Forum, 2002

Texture mapping is a common operation to increase the realism of three-dimensional meshes at low cost. We propose a new texture optimization algorithm based on the reduction of the physical space allotted to the texture image. Our algorithm optimizes the use of texture space by computing a warping function for the image and new texture coordinates. Neither the mesh geometry nor its connectivity are modified by the optimization. Our method uniformly distributes frequency content of the image in the spatial domain. In other words, the image is stretched in high frequency areas, whereas low frequency regions are shrunk. We also take into account distortions introduced by the mapping onto the model geometry in this process. The resulting image can be resampled at lower rate while preserving its original details. The unwarping is performed by the texture mapping function. Hence, the spaceoptimized texture is stored as-is in texture memory and is fully supported by current graphics hardware. We present several examples showing that our method significantly decreases texture memory usage without noticeable loss in visual quality.

View PDFchevron_right
Linear efficient antialiased displacement and reflectance mapping
Pierre Poulin

ACM Transactions on Graphics, 2013

We present Linear Efficient Antialiased Displacement and Reflectance (LEADR) mapping, a reflectance filtering technique for displacement mapped surfaces. Similarly to LEAN mapping, it employs two mipmapped texture maps, which store the first two moments of the displacement gradients. During rendering, the projection of this data over a pixel is used to compute a noncentered anisotropic Beckmann distribution using only simple, linear filtering operations. The distribution is then injected in a new, physically based, rough surface microfacet BRDF model, that includes masking and shadowing effects for both diffuse and specular reflection under directional, point, and environment lighting. Furthermore, our method is compatible with animation and deformation, making it extremely general and flexible. Combined with an adaptive meshing scheme, LEADR mapping provides the very first seamless and hardware-accelerated multi-resolution representation for surfaces. In order to demonstrate its ef...

View PDFchevron_right
Exact antialiasing of textured terrain models
Daniel Cohenor

The Visual Computer, 1997

We introduce a fast area-sampling antialiasing technique for textured terrain models. We scan the model in image order, averaging the pixel footprint values. The technique samples all, but only, the visible parts. It improves previous texture mapping methods that ignore selfoccluded footprints. It is superior to super-sampling. The image quality is assessed in both spatial and temporal domains. The good precision of the sampling process in the spatial domain provides an alias-free temporal domain. The low computational cost of the rendering technique and its high-quality filtering in the spatiotemporal domain offer a tool for real-time rendering of discrete terrain models.

View PDFchevron_right

Related topics

  • Algorithms
  • Computer Graphics
  • Three Dimensional Imaging
  • Information Storage and Retrieval
  • Image Enhancement
  • Reproducibility of Results
  • Image Texture
  • image resolution
    • 580 California St., Suite 400
    San Francisco, CA, 94104
    © 2025 Academia. All rights reserved