DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
The amendment filed January 30, 2026 has been entered. Claims 1-20 remain pending in the application. Applicant’s amendments to the Claims have overcome each and every objection previously set forth in the Non-Final Office Action mailed October 20, 2025.
Response to Arguments
Applicant's arguments filed January 30, 2026 have been fully considered but they are not persuasive.
In response to applicant’s argument that there is no motivation to combine the texture compression technique of Beers with the displacement data in Thonat, the examiner recognizes that obviousness may be established by combining or modifying the teachings of the prior art to produce the claimed invention where there is some teaching, suggestion, or motivation to do so found either in the references themselves or in the knowledge generally available to one of ordinary skill in the art. See In re Fine, 837 F.2d 1071, 5 USPQ2d 1596 (Fed. Cir. 1988), In re Jones, 958 F.2d 347, 21 USPQ2d 1941 (Fed. Cir. 1992), and KSR International Co. v. Teleflex, Inc., 550 U.S. 398, 82 USPQ2d 1385 (2007). In this case, Beers acknowledges that the compression technique can be used with other texture data, specifically displacement data (Paragraph 3 in 2nd Col. of Page 4 – “The VQ compression approach naturally extends to other classes of texture maps such as bump maps, displacement maps and environment maps…our preliminary results indicate that VQ compression works well for them”). Therefore, the embodiment of Beers that applies the compression technique to displacement data is used in the rejection. Additionally, one of ordinary skill in the art would be motivated to use Beers’ compression technique with the displacement data in Thonat for the benefits that come with compression, such as reducing memory usage (Beers: Paragraph 1 in 2nd Col. of Page 1 – “One way to alleviate these memory limitations is to store compressed representations”).
Moreover, the applicant argues that Beers discloses encoding a block of four codewords to a single codeword, and thus does not teach encoding a pair of values as a single value. In response to applicant's arguments against the references individually, one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986). Thonat teaches a pair of values from two separate arrays that is stored in every pixel/texel (Paragraph 2 in 2nd Col. of Page 5 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel”; Note: the minimum values of the mipmap form one array, and the maximum values of the mipmap form another array), and Beers teaches encoding pixels to a single value/codeword (Paragraph 2 in 2nd Col. of Page 2 – “Once the codebook has been generated we encode a texture by mapping each block of pixels to the nearest codeword…With Tree Structured VQ, we can traverse the code book tree always taking the path with the closest codeword”). While Beers does not teach exactly a pair of values being encoded, the reference does teach multiple values being encoded, such as the 2 x 2 codewords, which is interpreted as multiple pairs of values. Therefore, when Thonat and Beers are combined, they teach that the pair of values in the pixels (taught by Thonat) become encoded into a single pixel value (taught by Beers). The pairs of values are encoded to a codeword that they most closely resemble. In other words, the encoding technique from Beers can be used on the pair of values in Thonat.
Finally, the applicant argues that if Thonat and Beers were combined, a person of ordinary skill in the art would compress the upper and lower bound data into two separate codebooks, which does not meet the requirement of the claimed invention. However, Beers does not suggest having separate codebooks for different data (Paragraph 4 in 2nd Col. of Page 2 – “The size of a vector is dictated by the dimensions of the block of pixels being coded and the number of color channels used to define the color of each pixel. We can either design a codebook for each color channel separately, or treat components of a color as a single value and code them together. We use the latter approach, which results in a higher compression rate since only one codebook and index map is used, instead of one codebook and index map per color channel”). Therefore, when combining Thonat and Beers, a person of ordinary skill in the art would have one codebook even if there are multiple channels of data.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-4, 6-8, 10-12, and 14-19 are rejected under 35 U.S.C. 103 as being unpatentable over Thonat et al. (Tessellation-Free Displacement Mapping for Ray Tracing) in view of Beers et al. (Rendering from Compressed Textures), hereinafter Thonat and Beers respectively.
Regarding claim 1, Thonat teaches a method of compressing data for representing displacement information in a ray tracing system, wherein the displacement information indicates displacements to be applied to geometry in a scene to be rendered by the ray tracing system (Fig. 1, Paragraph 2 in 1st Col. of Page 4 – “Our method provides a direct ray tracing operator with low memory footprint for surfaces enriched with displacement maps. The key idea is to define a displacement-centric acceleration structure that can be mapped onto polygon meshes similarly as texture mapping”; Note: Fig. 1 shows a scene rendered using displacement maps), the method comprising:
retrieving a pair of datasets representing the displacement information, wherein a first of the datasets comprises a first array of values, and a second of the datasets comprises a second array of values (Paragraph 2 in 2nd Col. of Page 4 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel… This minmax mipmap is effectively a 2-channel texture with the same dimensions and same number of mipmap levels as the input displacement map, and is independent from the base surface on which the displacement is mapped on”; Note: the min values in the mipmap level is equivalent to the first array, and the max values in the mipmap level is equivalent to the second array);
and retrieving values from a corresponding array position in each of the first and second arrays, wherein the retrieved values form a pair of values representing an upper and lower bound of a magnitude of displacement for the corresponding array position (Paragraph 2 in 2nd Col. of Page 4, Paragraph 1-2 in 1st Col. of Page 5 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel… Given a ray and a base triangle, our D-BVH traversal visits all texels, at the finest mip level…The mipmap defines an implicit graph structure, where nodes are simply texels (2 integer coordinates and 1 integer mip level), and where children of a node are simply its 4 corresponding texels in the mip level below”; Note: texels are traversed and each texel contains minimum and maximum values of displacement).
Thonat does not teach identifying which of a plurality of predetermined conditions the pair of values satisfies; and encoding the pair of values as a single value in a compressed dataset, wherein the single value represents the identified predetermined condition. However, Beer teaches identifying which of a plurality of predetermined conditions the pair of values satisfies (Paragraph 2 in 2nd Col. of Page 2, Paragraph 5 in 1st Col. of Page 3 – “Once the codebook has been generated we encode a texture by mapping each block of pixels to the nearest codeword…To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords”; Note: the codebook comprises of predetermined conditions called codewords, and codewords, that satisfy or are closest to values in the mipmap, are identified); and encoding the pair of values as a single value in a compressed dataset, wherein the single value represents the identified predetermined condition (Paragraph 2 in 2nd Col. of Page 2, Paragraph 5 in 1st Col. of Page 3 – “Once the codebook has been generated we encode a texture by mapping each block of pixels to the nearest codeword…To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords”; Note: values in the mipmap are encoded into a single pixel codeword, and the codeword is a predetermined condition). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to identify a predetermined condition satisfied by the pair of values and encode the pair of values as a single value representing the predetermined condition because grouping multiple similar values into a single value helps reduce storage costs while being able to encompass the general information for later retrieval and use.
Regarding claim 2, Thonat in view of Beers teaches the method according to claim 1. Thonat does not teach repeatedly performing the encoding for all pairs of values comprised in the pair of datasets, wherein the compressed dataset comprises compressed displacement information for all information in the pair of datasets. However, Beers teaches repeatedly performing the encoding for all pairs of values comprised in the pair of datasets (Paragraph 4-5 in 1st Col. of Page 3 – “we can take advantage of the correlation between successive layers of a mipmap by encoding several levels at once, generating one codebook as well as one index map for the group of levels…The encoding of three mipmap levels is shown pictorially in figure 2. The full mipmap can be formed by repeating this process for every group of three levels, creating a separate index map and codebook for every group of three”), wherein the compressed dataset comprises compressed displacement information for all information in the pair of datasets (Paragraph 3 and 5 in 1st Col. of Page 3, Paragraph 3 in 2nd Col. of Page 4 – “A mipmap stores a texture as an image pyramid, and is designed to allow efficient filtering of a texture. Each mipmap level stores a filtered version of the texture corresponding to a particular image pixel to texture pixel ratio…To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords… The VQ compression approach naturally extends to other classes of texture maps such as bump maps, displacement maps and environment maps”; Note: the compressed mipmap contains texture data, which can be displacement data). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to encode all pairs of values for the benefit of a complete compression process, which helps reduce storage costs and increase loading efficiency. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to have the compressed dataset comprise compressed displacement information because “using less memory for textures may yield caching benefits, especially in cases where the textures do not fit in main memory and cause the machine to swap. One way to alleviate these memory limitations is to store compressed representations of textures in memory. A modified renderer could then render directly from this compressed representation” (Beers: Paragraph 4 in 1st Col. of Page 1, Paragraph 1 in 2nd Col. of Page 1). In other words, the size of the texture information, which includes displacement information, can be large, making it inefficient to load and render. Therefore, compressing texture information would help with being able to quickly load and render the data.
Regarding claim 3, Thonat in view of Beers teaches the method according to claim 1. Thonat further teaches wherein the pair of datasets representing displacement information represents a mipmap having been sampled from texel data representing a texture to be applied to the geometry in the scene (Paragraph 2 in 2nd Col. of Page 4 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel. More specifically, given a sampling ℎ of a displacement map with maximum resolution 𝑊 ×𝐻 and 𝐾 mip levels, we define its minmax mipmap…This minmax mipmap is effectively a 2-channel texture”; Note: the values in the mipmap correspond to texture data, and the mipmap was sampled from texel data of a displacement map).
Regarding claim 4, Thonat in view of Beers teaches the method according to claim 1. Thonat does not teach wherein the encoded single value representing the identified predetermined condition is represented with fewer bits than each value of the pair of values. However, Beers teaches wherein the encoded single value representing the identified predetermined condition is represented with fewer bits than each value of the pair of values (Paragraph 2 in 2nd Col. of Page 2, Paragraph 5 in 1st Col. of Page 3 – “Once the codebook has been generated we encode a texture by mapping each block of pixels to the nearest codeword…To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords”; Note: values in the mipmap are encoded into a single pixel codeword, and the codeword is a predetermined condition. A single pixel codeword has fewer bits than a 2 x 2 codeword). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to have the encoded single value contain less bits than the individual mipmap values for the benefit of reducing storage costs and increasing loading efficiency.
Regarding claim 6, Thonat in view of Beers teaches the method according to claim 1. Thonat does not teach compressing a plurality of pairs of datasets representing displacement information to form a plurality of compressed datasets, wherein each compressed dataset of the plurality of compressed datasets represents a different level of texture detail in the scene to be rendered. However, Beers teaches compressing a plurality of pairs of datasets representing displacement information to form a plurality of compressed datasets (Paragraph 5 in 1st Col. of Page 3, Paragraph 3 in 2nd Col. of Page 4 – “To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords… The VQ compression approach naturally extends to other classes of texture maps such as bump maps, displacement maps and environment maps”; Note: Thonat previously taught that the mipmap comprises min/max pairs of datasets in the rejection of claim 1. In this case, the mipmap is compressed and can contain displacement data), wherein each compressed dataset of the plurality of compressed datasets represents a different level of texture detail in the scene to be rendered (Paragraph 5 in 2nd Col. of Page 2, Paragraph 3 and 5 in 1st col. of Page 3 – “The size of the codebook influences the compression rate…we could use larger blocksizes to gain higher compression rates for worse quality…A mipmap stores a texture as an image pyramid, and is designed to allow efficient filtering of a texture. Each mipmap level stores a filtered version of the texture corresponding to a particular image pixel to texture pixel ratio… To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords”; Note: each level of mipmap represents a different level of texture detail since each level has a different codebook that affects texture quality). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to have the compressed dataset represent displacement information because “using less memory for textures may yield caching benefits, especially in cases where the textures do not fit in main memory and cause the machine to swap. One way to alleviate these memory limitations is to store compressed representations of textures in memory. A modified renderer could then render directly from this compressed representation” (Beers: Paragraph 4 in 1st Col. of Page 1, Paragraph 1 in 2nd Col. of Page 1). In other words, the size of the texture information, which includes displacement information, can be large, making it inefficient to load and render. Therefore, compressing texture information would help with being able to quickly load and render the data. It also would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to have each compressed dataset correspond to a different level of detail for the benefit of being able to prioritize the details of visible regions over non-visible regions.
Regarding claim 7, Thonat in view of Beers teaches the method according to claim 6. Thonat does not teach wherein each cell of each compressed dataset is generated in dependence on a plurality of regionally corresponding cells of a compressed dataset having a finer level of texture detail. However, Beers teaches wherein each cell of each compressed dataset is generated in dependence on a plurality of regionally corresponding cells of a compressed dataset having a finer level of texture detail (Paragraph 5 in 2nd Col. of Page 2, Paragraph 3 and 5 in 1st col. of Page 3 – “The size of the codebook influences the compression rate…we could use larger blocksizes to gain higher compression rates for worse quality…A mipmap stores a texture as an image pyramid, and is designed to allow efficient filtering of a texture. Each mipmap level stores a filtered version of the texture corresponding to a particular image pixel to texture pixel ratio… To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords”; Note: each level of mipmap represents a different level of texture detail since each level has a different codebook that affects texture quality). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to have the cells of the compressed dataset depend on corresponding cells of a compressed dataset with finer details because compressed data is, by definition, generated from less compressed data. In lossy compression, like the vector quantization method used in Beers, less compressed data contains more details.
Regarding claim 8, Thonat in view of Beers teaches the method according to claim 6. Thonat further teaches wherein the level of texture detail for each compressed dataset corresponds to a level of geometric detail (Paragraph 1 in 2nd Col. of Page 2, Paragraph 2 in 2nd Col. of Page 4, Fig. 5 Caption on Page 7 – “A flexible framework making possible to use customized intersection tests over the canonical base triangle mesh/displacement map input pair, yielding a wide spectrum of geometric fidelity levels with continuous level-of-details. Natural level-of-detail (LoD) support that can reduce geometric aliasing effects during rendering and improve performance…we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel… This minmax mipmap is effectively a 2-channel texture…Continuous geometric LoD: (a) Displaced object rendered at maximum LoD”; Note: mipmap texture levels correspond to geometric levels of detail).
Regarding claim 10, Thonat in view of Beers teaches the method according to claim 1. Thonat does not teach wherein the predetermined condition is that the pair of values is encompassed by a predetermined range of values, and wherein the encoded single value represents the predetermined range. However, Beers teaches wherein the predetermined condition is that the pair of values is encompassed by a predetermined range of values (Paragraph 4 in 1st Col. of Page 2, Paragraph 2 in 2nd Col. of Page 2 – “Each training vector is grouped with the nearest codeword, based on some distortion measure such as Euclidean distance… Once the codebook has been generated we encode a texture by mapping each block of pixels to the nearest codeword”; Note: the codebook comprises of predetermined conditions, which allows pixel values to be mapped to codewords. Codeword mappings are implied to be predetermined ranges of values in this case because pixel values must be within a certain distance of the codeword to be mapped to it. The distance is a range. For instance, if a pixel value is not close enough to a certain codeword, it would be mapped to another, closer codeword instead), and wherein the encoded single value represents the predetermined range (Paragraph 2 in 2nd Col. of Page 2, Paragraph 5 in 1st Col. of Page 3 – “Once the codebook has been generated we encode a texture by mapping each block of pixels to the nearest codeword…a codebook for the third level of the mipmap can be formed by averaging these 2 2 codewords down to single pixel codewords”; Note: the single pixel codeword represents a predetermined range since only those values closest to the codeword can be mapped to it). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to have the predetermined condition encompass a range and have the encoded single value represent the range for the benefit of being able to group similar data and replace them with a shorter value, which would help reduce storage costs.
Regarding claim 11, Thonat in view of Beers teaches the method according to claim 1. Thonat further teaches wherein one or more values available to encode the pair of values as a single value is reserved to represent a property of a texture to be applied to the geometry in the scene (Fig. 1, Paragraph 2 in 2nd Col. of Page 4 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel…This minmax mipmap is effectively a 2-channel texture”; Note: the values in the mipmap correspond to texture data. Fig. 1 shows how the texture is applied to the scene; see screenshot of Fig. 1 below).
PNG
media_image1.png
305
670
media_image1.png
Greyscale
Screenshot of Fig. 1 (taken from Thonat)
Regarding claim 12, Thonat in view of Beers teaches the method according to claim 1. Thonat further teaches wherein the first of the datasets comprises the first array of values representing minimum displacements, and the second of the datasets comprises the second array of values representing maximum displacements, wherein the upper and lower bound of a magnitude of displacement is represented by the pair of values as a respective maximum and minimum displacement value (Paragraph 2 in 2nd Col. of Page 4 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel… This minmax mipmap is effectively a 2-channel texture with the same dimensions and same number of mipmap levels as the input displacement map, and is independent from the base surface on which the displacement is mapped on”; Note: the min values in the mipmap level is equivalent to the first array, and the max values in the mipmap level is equivalent to the second array).
Regarding claim 14, Thonat in view of Beers teaches the method according to claim 1. Thonat further teaches wherein the upper and lower bound of a magnitude of displacement are configured to define boundaries of a bounding volume, associated with a respective position of the scene, to be used by the ray tracing system (Paragraph 2 in 2nd Col. of Page 4, Paragraph 2 in 2nd Col. of Page 5 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel…At each traversed texel, a 3D axis-aligned bounding box is computed by combining bounds from the D-BVH and information from the base triangle”).
Regarding claim 15, Thonat in view of Beers teaches the method according to claim 1. Thonat further teaches wherein the pair of values represents the upper and lower bound associated with the predetermined range of values, and relates to the upper and lower bound of the magnitude of displacement for the corresponding array position (Paragraph 2 in 2nd Col. of Page 4 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel. More specifically, given a sampling ℎ of a displacement map with maximum resolution𝑊 ×𝐻 and 𝐾 mip levels, we define its minmax mipmap”; Note: the minimum and maximum values of displacement are pairs of values representing the lower and upper bound of displacement). Thonat does not teach decompressing the compressed dataset by: retrieving the compressed dataset comprising an array of values, wherein each value of the array of values is an encoded value representing one of the plurality of predetermined conditions; retrieving a value from the compressed dataset; selecting, in dependence on the retrieved value, one of the plurality of predetermined conditions; in response to determining that the selected predetermined condition represents a predetermined range of values: generating a decompressed pair of values. However, Beers teaches decompressing the compressed dataset by: retrieving the compressed dataset comprising an array of values, wherein each value of the array of values is an encoded value representing one of the plurality of predetermined conditions (Paragraph 1 in 1st Col. of Page 2 – “A lossy–compressed version of the original image is represented as a set of indices into this codebook, with one index per block of pixels. This set of indices is called the index map. The texture can be decompressed by looking up each block of pixels in the codebook via its index”; Note: the set of indices is an array of values, and the indices are encoded values representing codewords, which are predetermined conditions); retrieving a value from the compressed dataset (Paragraph 1 in 1st Col. of Page 2 – “A lossy–compressed version of the original image is represented as a set of indices into this codebook, with one index per block of pixels. This set of indices is called the index map. The texture can be decompressed by looking up each block of pixels in the codebook via its index”; Note: indices are retrieved); selecting, in dependence on the retrieved value, one of the plurality of predetermined conditions (Paragraph 1 in 1st Col. of Page 2 – “A lossy–compressed version of the original image is represented as a set of indices into this codebook, with one index per block of pixels. This set of indices is called the index map. The texture can be decompressed by looking up each block of pixels in the codebook via its index”; Note: codewords are selected using indices); in response to determining that the selected predetermined condition represents a predetermined range of values: generating a decompressed pair of values (Paragraph 1 in 1st Col. of Page 2, Paragraph 2 in 2nd Col. Of Page 2, Paragraph 4 in 1st Col. of page 3 – “A lossy–compressed version of the original image is represented as a set of indices into this codebook, with one index per block of pixels. This set of indices is called the index map. The texture can be decompressed by looking up each block of pixels in the codebook via its index… Once the codebook has been generated we encode a texture by mapping each block of pixels to the nearest codeword… To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords”; Note: codewords are selected using indices. The codewords map back to mipmap blocks. The mipmap blocks comprise min/max pair values, as previously taught by Thonat in the rejection of claim 1). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to retrieve values from a compressed dataset, select a condition, and generate a pair of values because it provides a simple way to access some of the original data, and it is a common decompression technique used with vector quantization. Additionally, decompression helps reduce storage costs and increase loading efficiency.
Regarding claim 16, Thonat in view of Beers teaches the method according to claim 15. Thonat further teaches wherein a plurality of generated decompressed pairs forms a pair of decompressed datasets representing the displacement information, wherein a first of the decompressed datasets comprises a first array of values and a second of the decompressed datasets comprises a second array of values (Paragraph 2 in 2nd Col. of Page 4 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel. More specifically, given a sampling ℎ of a displacement map with maximum resolution 𝑊 ×𝐻 and 𝐾 mip levels, we define its minmax mipmap”; Note: the min part of the mipmap is equivalent to the first dataset, and the max part of the mipmap is equivalent to the second dataset. The compression and decompression of the datasets was previously taught by Beers). Thonat does not teach repeatedly retrieving and decompressing values comprised in the compressed dataset to obtain a plurality of generated decompressed pairs of values. Moreover, Beers does not directly teach repeatedly retrieving and decompressing values comprised in the compressed dataset to obtain a plurality of generated decompressed pairs of values. Instead, Beers separately teaches repeatedly retrieving and compressing values (Paragraph 4-5 in 1st Col. of Page 3 – “we can take advantage of the correlation between successive layers of a mipmap by encoding several levels at once, generating one codebook as well as one index map for the group of levels…The encoding of three mipmap levels is shown pictorially in figure 2. The full mipmap can be formed by repeating this process for every group of three levels, creating a separate index map and codebook for every group of three”), and decompressing values comprised in the compressed dataset to obtain a plurality of generated decompressed pairs of values (Paragraph 1 in 1st Col. of Page 2 – “A lossy–compressed version of the original image is represented as a set of indices into this codebook, with one index per block of pixels. This set of indices is called the index map. The texture can be decompressed by looking up each block of pixels in the codebook via its index”). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the features of repeating operations and decompressing values comprised in the compressed dataset to repeatedly retrieve and decompress values because decompression is a common process that follows compression since in many cases, compressed data would need to be decompressed in order to be used. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to repeatedly retrieve and decompress values to generate pairs for the benefit of being able to obtain and use the original data.
Regarding claim 17, Thonat in view of Beers teaches the method according to claim 16. Thonat further teaches wherein the first array of values represents minimum displacements, and the second array of values represents maximum displacements, and wherein each generated decompressed pair represents a respective maximum and minimum displacement value (Paragraph 2 in 2nd Col. of Page 4 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel… This minmax mipmap is effectively a 2-channel texture with the same dimensions and same number of mipmap levels as the input displacement map, and is independent from the base surface on which the displacement is mapped on”; Note: the min values in the mipmap is equivalent to the first array, and the max values in the mipmap is equivalent to the second array. Additionally, the feature that the pairs are decompressed and generated was previously taught by Beers in the rejection of claim 15 above).
Regarding claim 18, Thonat in view of Beers teaches the method according to claim 15. Thonat further teaches wherein the generated decompressed pair of values representing the upper and lower bound of a magnitude of displacement are configured to define boundaries of a bounding volume, associated with a respective position of the scene, to be used by the ray tracing system (Paragraph 2 in 2nd Col. of Page 4, Paragraph 2 in 2nd Col. of Page 5 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel…At each traversed texel, a 3D axis-aligned bounding box is computed by combining bounds from the D-BVH and information from the base triangle”; Note: the feature that the pairs are decompressed and generated was previously taught by Beers in the rejection of claim 15 above).
Regarding claim 19, Thonat teaches a hardware unit for use in a ray tracing system (Paragraph 6 in 2nd Col. of Page 8 – “We implemented our method on GPU using the Vulkan Ray Tracing extension [Khronos 2020], but it is applicable to any ray tracing system supporting custom geometries”), wherein the hardware unit is configured to:
retrieve a pair of datasets representing the displacement information, wherein a first of the datasets comprises a first array of values, and a second of the datasets comprises a second array of values (Paragraph 2 in 2nd Col. of Page 4 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel… This minmax mipmap is effectively a 2-channel texture with the same dimensions and same number of mipmap levels as the input displacement map, and is independent from the base surface on which the displacement is mapped on”; Note: the min values in the mipmap level is equivalent to the first array, and the max values in the mipmap level is equivalent to the second array);
and retrieve values from a corresponding array position in each of the first and second arrays, wherein the retrieved values form a pair of values representing an upper and lower bound of a magnitude of displacement for the corresponding array position (Paragraph 2 in 2nd Col. of Page 4, Paragraph 1-2 in 1st Col. of Page 5 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel… Given a ray and a base triangle, our D-BVH traversal visits all texels, at the finest mip level…The mipmap defines an implicit graph structure, where nodes are simply texels (2 integer coordinates and 1 integer mip level), and where children of a node are simply its 4 corresponding texels in the mip level below”; Note: texels are traversed and each texel contains minimum and maximum values of displacement).
Thonat does not teach identifying which of a plurality of predetermined conditions the pair of values satisfies; and encoding the pair of values as a single value in a compressed dataset, wherein the single value represents the identified predetermined condition. However, Beer teaches identifying which of a plurality of predetermined conditions the pair of values satisfies (Paragraph 2 in 2nd Col. of Page 2, Paragraph 5 in 1st Col. of Page 3 – “Once the codebook has been generated we encode a texture by mapping each block of pixels to the nearest codeword…To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords”; Note: the codebook comprises of predetermined conditions called codewords, and codewords, that satisfy or are closest to values in the mipmap, are identified); and encoding the pair of values as a single value in a compressed dataset, wherein the single value represents the identified predetermined condition (Paragraph 2 in 2nd Col. of Page 2, Paragraph 5 in 1st Col. of Page 3 – “Once the codebook has been generated we encode a texture by mapping each block of pixels to the nearest codeword…To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords”; Note: values in the mipmap are encoded into a single pixel codeword, and the codeword is a predetermined condition). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to identify a predetermined condition satisfied by the pair of values and encode the pair of values as a single value representing the predetermined condition because grouping multiple similar values into a single value helps reduce storage costs while being able to encompass the general information for later retrieval and use.
Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Thonat in view of Beers and Barman et al. (A quantization based codebook formation method of vector quantization algorithm to improve the compression ratio while preserving the visual quality of the decompressed image), hereinafter Barman.
Regarding claim 5, Thonat in view of Beers teaches the method according to claim 4. Thonat does not teach wherein the encoded single value is represented using 2 or 3 bits. However, Barman teaches wherein the encoded single value is represented using 2 bits (Paragraph 1 on Page 1 – “the second group’s codewords are quantized into two bit values, namely 0, 1, 2, and 3, to increase compression ratio”). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Barman to have the encoded single value be 2 bits for the benefit of further compression since having 2-bit codewords implies there are not many codewords for values to be mapped to. Additionally, in cases where there are not many values, having several codewords would be excessive and/or result in less compression.
Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Thonat in view of Beers and Lier et al. (A High-Resolution Compression Scheme for Ray Tracing Subdivision Surfaces with Displacement), hereinafter Lier.
Regarding claim 9, Thonat in view of Beers teaches the method according to claim 1. Thonat does not teach decoding the compressed dataset to obtain at least one pair of values representing an upper and lower bound of a magnitude of displacement, wherein the compressed dataset is associated with a patch within the scene; determining a bounding volume that contains the patch, wherein at least one dimension of the bounding volume is determined in dependence on the pair of values; and responsive to determining that the ray intersects the bounding volume, in dependence on tessellation indications associated with the patch, subdividing the patch one or more times to obtain a plurality of patch sub-units. However, Beers teaches decoding the compressed dataset to obtain at least one pair of values representing an upper and lower bound of a magnitude of displacement (Paragraph 1 in 1st Col. of Page 2 – “A lossy–compressed version of the original image is represented as a set of indices into this codebook, with one index per block of pixels. This set of indices is called the index map. The texture can be decompressed by looking up each block of pixels in the codebook via its index”; Note: the compressed data is decoded to obtain codewords, which correspond to values. Additionally, the pair of values representing an upper and lower bound of displacement was previously taught by Thonat in the rejection of claim 1 above). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to decode the compressed dataset for the benefit of being able to obtain and directly use the original data. Thonat modified by Beers still does not teach wherein the compressed dataset is associated with a patch within the scene; determining a bounding volume that contains the patch, wherein at least one dimension of the bounding volume is determined in dependence on the pair of values; and responsive to determining that the ray intersects the bounding volume, in dependence on tessellation indications associated with the patch, subdividing the patch one or more times to obtain a plurality of patch sub-units. However, Lier teaches wherein the compressed dataset is associated with a patch within the scene (Fig. 2 Caption on Page 5 – “For almost flat sub-patches, compressed and quantized CBVHs are built and embedded in the standard BVH. The left panel shows a whole patch and its surface vertices”; Note: Fig. 2 shows a patch of a compressed bounding volume hierarchy (CBVH); see screenshot of Fig. 2 below); determining a bounding volume that contains the patch, wherein at least one dimension of the bounding volume is determined in dependence on the pair of values (Paragraph 4 on Page 4, Fig. 2 Caption on Page 5 – “The first element of the compression is an aggressive hierarchical quantization, which stores bounds of subnodes relative to their parent nodes and requires only few bits… As it can be seen in Figure 3, directly adjacent bounding volumes have very similar bounds…For almost flat sub-patches, compressed and quantized CBVHs are built and embedded in the standard BVH. The left panel shows a whole patch and its surface vertices. As depicted right, the root of a CBVH is aligned along a subset of vertices of the whole patch and all child-nodes are understood to be in this CBVH-specific frame of reference”; Note: Fig. 2 shows a bounding volume containing a patch. The bounding volume is generated in dependence of bounds); and responsive to determining that the ray intersects the bounding volume (Paragraph 1 on Page 9 – “we use an approximate intersection test that projects the ray onto a straight line on the surface. The line is constructed from the entry and exit points of the bounding box projected to surface points (see Figure 9)”), in dependence on tessellation indications associated with the patch, subdividing the patch one or more times to obtain a plurality of patch sub-units (Paragraph 1 on Page 4, Paragraph 1 on Page 6 – “The top-level is a standard, uncompressed BVH composed of globally axis-aligned bounding volumes that contain the scene’s subdivided patches. These patch fragments, or subpatches, must satisfy a certain criterion in flatness, which can be determined with the opening angle of the surface’s normals. The bottom-level, and thus the bulk of the hierarchy, is then made up of compressed and quantized 4-wide BVHs (called CBVHs)…we will introduce a custom-made, fast intersection test for our surface approximations with only little demand in extra computation”; Note: patches are subdivided based on flatness, which is a tessellation indication).
PNG
media_image2.png
191
925
media_image2.png
Greyscale
Screenshot of Fig. 2 (taken from Beers)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Lier to associate the compressed dataset with a patch and determine a bounding volume that contains the patch for the benefit of identifying which area in the scene corresponds to the compressed dataset, which is important for when the compressed data needs to be accessed later for decompression or processing. It also would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Lier to subdivide the patch in response to determine that the ray intersects the bounding volume because the ray intersection indicates that the region is visible, and in order to clearly render or display the region to a user, it should be decompressed to show the details.
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over Thonat in view of Beers and An et al. (Bi-dimensional empirical mode decomposition (BEMD) algorithm based on particle swarm optimization-fractal interpolation), hereinafter An.
Regarding claim 13, Thonat in view of Beers teaches the method according to claim 1. Thonat further teaches wherein the upper and lower bound of a magnitude of displacement is represented by the pair of values (Paragraph 2 in 2nd Col. of Page 4 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel… This minmax mipmap is effectively a 2-channel texture with the same dimensions and same number of mipmap levels as the input displacement map, and is independent from the base surface on which the displacement is mapped on”; Note: the min values in the mipmap level is equivalent to the first array, and the max values in the mipmap level is equivalent to the second array). Thonat does not teach wherein the first of the datasets comprises the first array of values of mid-point displacements, and the second of the datasets comprises the second array of values representing halfwidth displacements. However, An teaches wherein the first of the datasets comprises the first array of values of mid-point displacements, and the second of the datasets comprises a second array of values representing halfwidth displacements (Fig. 3 on Page 10 – The modified screenshot of Fig. 3 below shows midpoint and halfwidth displacements. Additionally, the first and second arrays of values were previously taught by Thornat in the rejection of claim 1 above).
PNG
media_image3.png
821
823
media_image3.png
Greyscale
Modified screenshot of Fig. 3 (taken from An)
A person of ordinary skill in the art before the effective filing date of the claimed invention would have recognized that the minimum and maximum displacements of Thonat could have been substituted for the mid-point and halfwidth displacements of An because both the minimum/maximum displacements and mid-point/halfwidth displacements can serve the purpose of representing upper and lower bounds of displacement and are both types of displacement values. Furthermore, a person of ordinary skill in the art would have been able to carry out the substitution. Finally, the substitution achieves the predictable result of representing upper and lower bounds of displacement. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to substitute the minimum/maximum displacements of Thonat for the mid-point/halfwidth displacements of An according to known methods to yield the predictable result of representing upper and lower bounds of displacement.
Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Thonat in view of Beers and Peterson et al. (US 20180365879 A1), hereinafter Peterson.
Regarding claim 20, Thonat teaches a hardware unit (Paragraph 6 in 2nd Col. of Page 8 – “We implemented our method on GPU using the Vulkan Ray Tracing extension [Khronos 2020], but it is applicable to any ray tracing system supporting custom geometries”) configured to:
retrieve a pair of datasets representing the displacement information, wherein a first of the datasets comprises a first array of values, and a second of the datasets comprises a second array of values (Paragraph 2 in 2nd Col. of Page 4 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel… This minmax mipmap is effectively a 2-channel texture with the same dimensions and same number of mipmap levels as the input displacement map, and is independent from the base surface on which the displacement is mapped on”; Note: the min values in the mipmap level is equivalent to the first array, and the max values in the mipmap level is equivalent to the second array);
and retrieve values from a corresponding array position in each of the first and second arrays, wherein the retrieved values form a pair of values representing an upper and lower bound of a magnitude of displacement for the corresponding array position (Paragraph 2 in 2nd Col. of Page 4, Paragraph 1-2 in 1st Col. of Page 5 – “we use a minmax mipmap as an acceleration structure for the displacement, storing both a conservative minimum and maximum value of displacement over the 𝑢𝑣-domain corresponding to each texel… Given a ray and a base triangle, our D-BVH traversal visits all texels, at the finest mip level…The mipmap defines an implicit graph structure, where nodes are simply texels (2 integer coordinates and 1 integer mip level), and where children of a node are simply its 4 corresponding texels in the mip level below”; Note: texels are traversed and each texel contains minimum and maximum values of displacement).
Thonat does not teach a non-transitory computer readable storage medium having stored thereon a computer readable dataset description of an integrated circuit that, when processed in an integrated circuit manufacturing system, configures the integrated circuit manufacturing system to manufacture a hardware unit for use in a ray tracing system. However, Peterson teaches a non-transitory computer readable storage medium having stored thereon a computer readable dataset description of an integrated circuit that, when processed in an integrated circuit manufacturing system, configures the integrated circuit manufacturing system to manufacture a hardware unit for use in a ray tracing system (Paragraph 0061 – “The ray tracing systems described herein may be embodied in hardware on an integrated circuit. There may be provided a method of manufacturing, at an integrated circuit manufacturing system, a ray tracing system…There may be provided a non-transitory computer readable storage medium having stored thereon a computer readable description of an integrated circuit that, when processed in an integrated circuit manufacturing system, causes the integrated circuit manufacturing system to manufacture a ray tracing system as described herein”). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Peterson to have a non-transitory computer readable storage medium for the benefit of a reliable and persistent storage.
Thonat modified by Peterson still does not teach identifying which of a plurality of predetermined conditions the pair of values satisfies; and encoding the pair of values as a single value in a compressed dataset, wherein the single value represents the identified predetermined condition. However, Beer teaches identifying which of a plurality of predetermined conditions the pair of values satisfies (Paragraph 2 in 2nd Col. of Page 2, Paragraph 5 in 1st Col. of Page 3 – “Once the codebook has been generated we encode a texture by mapping each block of pixels to the nearest codeword…To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords”; Note: the codebook comprises of predetermined conditions called codewords, and codewords, that satisfy or are closest to values in the mipmap, are identified); and encoding the pair of values as a single value in a compressed dataset, wherein the single value represents the identified predetermined condition (Paragraph 2 in 2nd Col. of Page 2, Paragraph 5 in 1st Col. of Page 3 – “Once the codebook has been generated we encode a texture by mapping each block of pixels to the nearest codeword…To compress a mipmap, we begin by compressing the original texture using 4 x 4 blocks. Given the codebook for this first mipmap level, we can form a codebook for the next level by averaging each 4 x 4 codeword down to a 2 x 2 codeword. Similarly, a codebook for the third level of the mipmap can be formed by averaging these 2 x 2 codewords down to single pixel codewords”; Note: values in the mipmap are encoded into a single pixel codeword, and the codeword is a predetermined condition). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Thonat to incorporate the teachings of Beer to identify a predetermined condition satisfied by the pair of values and encode the pair of values as a single value representing the predetermined condition because grouping multiple similar values into a single value helps reduce storage costs while being able to encompass the general information for later retrieval and use.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Ali et al. (Compressed Facade Displacement Maps) teaches a method of rendering urban models from compressed displacement maps.
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MICHELLE HAU MA whose telephone number is (571)272-2187. The examiner can normally be reached M-Th 7-5:30.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, King Poon can be reached at (571) 270-0728. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MICHELLE HAU MA/ Examiner, Art Unit 2617
/KING Y POON/Supervisory Patent Examiner, Art Unit 2617