Prosecution Insights
Last updated: November 08, 2026
Application No. 19/007,895

Systems and Methods for Directly Generating Splats from Three-Dimensional Image Data

Non-Final OA §103
Filed
Jan 02, 2025
Examiner
NGUYEN, ANH TUAN V
Art Unit
2619
Tech Center
2600 — Communications
Assignee
Illuscio Inc.
OA Round
1 (Non-Final)
72%
Grant Probability
Favorable
1-2
OA Rounds
1y 0m
Est. Remaining
91%
With Interview

Examiner Intelligence

Grants 72% — above average
72%
Career Allowance Rate
369 granted / 512 resolved
+10.1% vs TC avg
Strong +19% interview lift
Without
With
+19.1%
Interview Lift
resolved cases with interview
Typical timeline
2y 10m
Avg Prosecution
25 currently pending
Career history
548
Total Applications
across all art units

Statute-Specific Performance

§101
9.2%
-30.8% vs TC avg
§103
69.4%
+29.4% vs TC avg
§102
4.6%
-35.4% vs TC avg
§112
12.4%
-27.6% vs TC avg
Black line = Tech Center average estimate • Based on career data from 512 resolved cases

Office Action

§103
DETAILED ACTION The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . 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 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. Claim Rejections - 35 USC § 103 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. Claim(s) 1, 7, 9, 11, 13, and 19-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lin et al. ("Direct learning of mesh and appearance via 3d gaussian splatting." arXiv preprint arXiv:2405.06945 (2024)) in view of Tromans (US 2026/0228886) and Tanner et al. (US 2025/0316029). Regarding claim 1, Lin teaches/suggests: A method comprising: receiving a plurality of three-dimensional (3D) primitives for a non-splat 3D representation of a 3D object (Lin Fig. 2 “The mesh is derived from a learnable SDF grid” §III-B ¶1 “given a point x in 3D space, the SDF value is defined as”); selecting a set of 3D primitives from the plurality of 3D primitives (Lin Fig. 2 “Gaussians are created from the mesh faces” [The created Gaussians meet the selecting.]); generating one or more splats that form a first surface that matches a second surface formed by the set of 3D primitives (Lin Fig. 2 “Gaussians are created from the mesh faces, ensuring their alignment with the surface” §III-A ¶1 “3D Gaussian Splatting (3DGS) [14] represents the 3D scene using a collection of Gaussians”); defining visual characteristics for the one or more splats (Lin Fig. 2 “A neural appearance model determines colors for Gaussians”); and presenting the one or more splats instead of the set of 3D primitives in response to a request to access the 3D object (Lin §IV-D ¶1 “the rendered Gaussians include both the Gaussians derived from the mesh and the additional background Gaussians” [The rendered Gaussians meet the request.] §V ¶1 “Note that the mesh is not used in the rendering of our method”). Lin does not teach/suggest: generating one or more splats that form a first surface that matches a second surface formed by the set of 3D primitives with an acceptable amount of structural loss; Tromans, however, teaches/suggests an acceptable amount of structural loss (Tromans [0112] “For the shape loss, the threshold is typically 3 pixels or 0.432 mm”). Before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to modify the Gaussians of Lin to be within a shape (i.e., structural) loss threshold as taught/suggested by Tromans for accuracy. As such, Lin as modified by Tromans teaches/suggests: generating one or more splats that form a first surface that matches a second surface formed by the set of 3D primitives with an acceptable amount of structural loss (Lin Fig. 2 “Gaussians are created from the mesh faces, ensuring their alignment with the surface” Tromans [0112] “For the shape loss, the threshold is typically 3 pixels or 0.432 mm”); Lin further teaches/suggests visual characteristic loss (Lin §IV-D ¶1 “Our method relies only on photometric supervision and utilizes the loss function ... for comparing the rendered image against the ground truth”). Lin as modified by Tromans does not teach/suggest: defining visual characteristics for the one or more splats to match visual characteristics across the second surface formed by the set of 3D primitives with an acceptable amount of visual characteristic loss; Tanner, in view of Lin, teaches/suggests: defining visual characteristics for the one or more splats to match visual characteristics across the second surface formed by the set of 3D primitives with an acceptable amount of visual characteristic loss (Lin Fig. 2 “A neural appearance model determines colors for Gaussians” §IV-D ¶1 “Our method relies only on photometric supervision and utilizes the loss function ... for comparing the rendered image against the ground truth” Tanner [0089] “comparing one or more visual features of the generated mesh to one or more visual features of the stored meshes. For example, process 700 can determine whether a threshold number of visual features in the generated mesh match a threshold number of visual features in a stored mesh, whether the generated mesh matches a stored mesh within a threshold amount of variation, etc.” [In view of Lin and Tanner, the visual features of the Gaussians would be compared to those of the mesh faces.]); Before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to modify the Gaussians of Lin as modified by Tromans to be within a visual loss threshold of the mesh faces as taught/suggested by Tanner for accuracy. Regarding claim 7, Lin as modified by Tromans and Tanner teaches/suggests: The method of claim 1 further comprising: orienting each of the one or more splats according to a surface normal associated with the second surface or one or more of the set of 3D primitives (Lin Fig. 2 “Gaussians are created from the mesh faces, ensuring their alignment with the surface” §II-B ¶3 “ensure consistency between the Gaussian orientation and the surface normal”). Regarding claim 9, Lin as modified by Tromans and Tanner teaches/suggests: The method of claim 1, wherein defining the visual characteristics comprises: determining that the one or more splats are insufficient to recreate the visual characteristics across the second surface with the acceptable amount of visual characteristic loss (Lin §V-F ¶4 “we choose 3 Gaussians per face during the joint learning of mesh and Gaussian appearance” Tanner [0089] “comparing one or more visual features of the generated mesh to one or more visual features of the stored meshes. For example, process 700 can determine whether a threshold number of visual features in the generated mesh match a threshold number of visual features in a stored mesh, whether the generated mesh matches a stored mesh within a threshold amount of variation, etc.”); generating a set of splats that form the first surface with more splats than the one or more splats (Lin §V-F ¶4 “use 6 Gaussians for the refinement stage”); and increasing visual characteristic variation across the first surface by defining the visual characteristics of the set of splats to match visual characteristics across the second surface formed by the set of 3D primitives with the acceptable amount of visual characteristic loss (Lin §V-F ¶4 “use 6 Gaussians for the refinement stage” Tanner [0089] “comparing one or more visual features of the generated mesh to one or more visual features of the stored meshes. For example, process 700 can determine whether a threshold number of visual features in the generated mesh match a threshold number of visual features in a stored mesh, whether the generated mesh matches a stored mesh within a threshold amount of variation, etc.”). The same rationale to combine as set forth in the rejection of claim 1 is incorporated herein. Regarding claim 11, Lin as modified by Tromans and Tanner teaches/suggests: The method of claim 1, wherein presenting the one or more splats comprises: loading the one or more splats instead of the set of 3D primitives into memory of a device that renders a visualization of the 3D object (Lin §IV-D ¶1 “the rendered Gaussians include both the Gaussians derived from the mesh and the additional background Gaussians” §V ¶1 “Note that the mesh is not used in the rendering of our method” §V-A ¶1 “Experiments are conducted on an RTX3090 GPU with 24GB memory”). Claims 13 and 19 recite limitation(s) similar in scope to those of claims 1 and 7, respectively, and are rejected for the same reason(s). Lin as modified by Tromans and Tanner further teaches/suggests one or more hardware processors (Lin §V-A ¶1 “Experiments are conducted on an RTX3090 GPU with 24GB memory”). Claim 20 recites limitation(s) similar in scope to those of claim 1, and is rejected for the same reason(s). Lin as modified by Tromans and Tanner further teaches/suggests a non-transitory computer-readable medium storing instructions (Lin §V-A ¶1 “Experiments are conducted on an RTX3090 GPU with 24GB memory”). Claim(s) 2-3 and 14-15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lin et al. ("Direct learning of mesh and appearance via 3d gaussian splatting." arXiv preprint arXiv:2405.06945 (2024)) in view of Tromans (US 2026/0228886) and Tanner et al. (US 2025/0316029) as applied to claims 1 and 13 above, and further in view of Seok (US 2012/0320152) and Siri (US 2019/0188523). Regarding claim 2, Lin as modified by Tromans and Tanner does not teach/suggest: The method of claim 1, wherein selecting the set of 3D primitives comprises: determining two or more 3D primitives from the plurality of 3D primitives that are positioned next to one another. Seok, however, teaches/suggests: determining two or more 3D primitives from the plurality of 3D primitives that are positioned next to one another (Seok [0024] “The adjacent triangles may be combined according to the comparison of the RGB levels and considered as a single plane”). Before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to modify the mesh of Lin as modified by Tromans and Tanner such that adjacent triangles are selected as taught/suggested by Seok to be combined. Lin as modified by Tromans, Tanner, and Seok does not teach/suggest that deviate from a common plane by less than a first threshold amount. Siri, however, teaches/suggests that deviate from a common plane by less than a first threshold amount (Siri [0212] “choosing a distance threshold from the single plane on which each triangle of object A lays with respect to the points of object B”). Before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to modify the selection of Lin as modified by Tromans, Tanner, and Seok to include the distance threshold of Siri to combine more triangles. Regarding claim 3, Lin as modified by Tromans, Tanner, Seok, and Siri teaches/suggests: The method of claim 2, wherein selecting the set of 3D primitives further comprises: determining color variation across the two or more 3D primitives is less than a second threshold amount (Seok [0024] “the maximum value among the RGB levels in one triangle is selected and compared with one value among the RGB levels corresponding to one value selected among the RGB levels in another triangle. When the two values are similar, it is determined as a single plane”). The same rationale to combine as set forth in the rejection of claim 2 is incorporated herein. Claims 14 and 15 recite limitation(s) similar in scope to those of claims 2 and 3, respectively, and are rejected for the same reason(s). Claim(s) 4 and 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lin et al. ("Direct learning of mesh and appearance via 3d gaussian splatting." arXiv preprint arXiv:2405.06945 (2024)) in view of Tromans (US 2026/0228886) and Tanner et al. (US 2025/0316029) as applied to claims 1 and 13 above, and further in view of Seok (US 2012/0320152). Regarding claim 4, Lin as modified by Tromans and Tanner does not teach/suggest: The method of claim 1, wherein generating the one or more splats comprises: defining fewer splats to form the first surface than 3D primitives in the set of 3D primitives that form the second surface. Seok, in view of Lin, teaches/suggests: defining fewer splats to form the first surface than 3D primitives in the set of 3D primitives that form the second surface (Lin Fig. 2 “Gaussians are created from the mesh faces” Seok [0024] “The adjacent triangles may be combined according to the comparison of the RGB levels and considered as a single plane” [In view of Lin and Seok, the created Gaussians for the combined triangles meet the defining.]). The same rationale to combine as set forth in the rejection of claim 2 is incorporated herein. Claim 16 recites limitation(s) similar in scope to those of claim 4, and is rejected for the same reason(s). Claim(s) 5 and 17 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lin et al. ("Direct learning of mesh and appearance via 3d gaussian splatting." arXiv preprint arXiv:2405.06945 (2024)) in view of Tromans (US 2026/0228886) and Tanner et al. (US 2025/0316029) as applied to claims 1 and 13 above, and further in view of Johnson et al. (US 2009/0067670). Regarding claim 5, Lin as modified by Tromans and Tanner teaches/suggests: The method of claim 1, wherein each splat of the one or more splats has a Lin as modified by Tromans and Tanner does not teach/suggest a varying ellipsoidal shape. Johnson, however, teaches/suggests a varying ellipsoidal shape (Johnson [0109] “An ellipse may be fitted to the blob by creating a series of ellipses of various sizes and comparing each with the blob to determine a best fit”). Before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to modify the Gaussians of Lin as modified by Tromans and Tanner such that they have varying sizes as taught/suggested by Johnson for best fit. Claim 17 recites limitation(s) similar in scope to those of claim 5, and is rejected for the same reason(s). Claim(s) 6 and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lin et al. ("Direct learning of mesh and appearance via 3d gaussian splatting." arXiv preprint arXiv:2405.06945 (2024)) in view of Tromans (US 2026/0228886) and Tanner et al. (US 2025/0316029) as applied to claims 1 and 13 above, and further in view of Johnson et al. (US 2009/0067670) and Seok (US 2012/0320152). Regarding claim 6, Lin as modified by Tromans and Tanner teaches/suggests: The method of claim 1, wherein each splat of the one or more splats is defined with a position in a 3D space and one or more radii (Lin §I ¶2 “3DGS employs an explicit representation called Gaussians, which represent anisotropic ellipsoids in 3D space”). Lin as modified by Tromans and Tanner does not teach/suggest that adjust a shape of the splat. Johnson, however, teaches/suggests that adjust a shape of the splat (Johnson [0109] “An ellipse may be fitted to the blob by creating a series of ellipses of various sizes and comparing each with the blob to determine a best fit”). The same rationale to combine as set forth in the rejection of claim 5 is incorporated herein. Lin as modified by Tromans, Tanner, and Johnson does not teach/suggest wherein each splat spans a region of the first surface that is larger than a region of the second surface spanned by a single 3D primitive of the set of 3D primitives. Seok, however, teaches/suggests a region of the first surface that is larger than a region of the second surface spanned by a single 3D primitive of the set of 3D primitives (Seok [0024] “The adjacent triangles may be combined according to the comparison of the RGB levels and considered as a single plane”). The combined triangles meet the region of the first surface. Before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to modify the mesh of Lin as modified by Tromans, Tanner, and Johnson such that adjacent triangles are selected as taught/suggested by Seok to be combined. As such, Lin as modified by Tromans, Tanner, Johnson, and Seok teaches/suggests each splat spans a region of the first surface that is larger than a region of the second surface spanned by a single 3D primitive of the set of 3D primitives (Johnson [0109] “An ellipse may be fitted to the blob by creating a series of ellipses of various sizes and comparing each with the blob to determine a best fit” Seok [0024] “The adjacent triangles may be combined according to the comparison of the RGB levels and considered as a single plane”). Claim 18 recites limitation(s) similar in scope to those of claim 6, and is rejected for the same reason(s). Claim(s) 8 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lin et al. ("Direct learning of mesh and appearance via 3d gaussian splatting." arXiv preprint arXiv:2405.06945 (2024)) in view of Tromans (US 2026/0228886) and Tanner et al. (US 2025/0316029) as applied to claim 1 above, and further in view of Popescu et al. (US 2005/0128196) and Reiner et al. (US 2014/0354637). Regarding claim 8, Lin as modified by Tromans and Tanner does not teach/suggest: The method of claim 1, wherein defining the visual characteristics comprises: receiving a texture that is associated with the non-splat 3D representation; mapping a set of visual characteristics from the texture to the second surface formed by the set of 3D primitives; and defining the visual characteristics of each splat of the one or more splats based on a subset of visual characteristics from the set of visual characteristics. Popescu, however, teaches/suggests: receiving a texture that is associated with the non-splat 3D representation (Popescu [0144] “The color data is merged into a cube map panorama”); mapping a set of visual characteristics from the texture to the second surface formed by the set of 3D primitives (Popescu [0149] “The 3D triangle mesh is texture-mapped with the cube map faces”); and defining the visual characteristics of each splat of the one or more splats based on Before the effective filing date of the claimed invention, the substitution of one known element (the cube map of Popescu) for another (the neural appearance model of Lin) would have been obvious to one of ordinary skill in the art because such substitutions would have yielded predictable results, namely, for texturing. As such, Lin as modified by Tromans, Tanner, and Popescu teaches/suggests: defining the visual characteristics of each splat of the one or more splats based on a subset of visual characteristics from the set of visual characteristics (Tanner [0089] “comparing one or more visual features of the generated mesh to one or more visual features of the stored meshes. For example, process 700 can determine whether a threshold number of visual features in the generated mesh match a threshold number of visual features in a stored mesh, whether the generated mesh matches a stored mesh within a threshold amount of variation, etc.” Popescu [0150] “The splats are texture mapped with the cube map faces” [In view of Tanner and Popescu, the threshold number of features meets the subset of visual characteristics.]). Popescu is silent regarding that are mapped to a portion of the second surface that aligns with a portion of the first surface spanned by that splat. Reiner, however, teaches/suggests that are mapped to a portion of the second surface that aligns with a portion of the first surface spanned by that splat (Reiner [0057] “The ellipsoidal shape is texture mapped by projecting the photometric information, for example, RGB color information (R, G and B channels) of the corresponding entity representative point over the ellipsoidal shape using a Gaussian splat”). Before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to modify the texture mapping of the Gaussians of Lin as modified by Tromans, Tanner, and Popescu to correspond to that of the mesh faces as taught/suggested by Reiner for accuracy. Claim(s) 10 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lin et al. ("Direct learning of mesh and appearance via 3d gaussian splatting." arXiv preprint arXiv:2405.06945 (2024)) in view of Tromans (US 2026/0228886) and Tanner et al. (US 2025/0316029) as applied to claim 1 above, and further in view of Swaminathan et al. (US 2018/0160160). Regarding claim 10, Lin as modified by Tromans and Tanner does not teach/suggest: The method of claim 1, wherein presenting the one or more splats comprises: streaming the one or more splats instead of the set of 3D primitives over a data network to a requesting client device. Swaminathan, however, teaches/suggests streaming (Swaminathan [0151] “streaming each tile to the client device”). Before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to modify the Gaussians of Lin as modified by Tromans and Tanner to be streamed as taught/suggested by Swaminathan for display. As such, Lin as modified by Tromans, Tanner, and Swaminathan teaches/suggests: streaming the one or more splats instead of the set of 3D primitives over a data network to a requesting client device (Lin §IV-D ¶1 “the rendered Gaussians include both the Gaussians derived from the mesh and the additional background Gaussians” §V ¶1 “Note that the mesh is not used in the rendering of our method” Swaminathan [0151] “streaming each tile to the client device” [0032] “The one or more server devices 102, one or more content provider devices 106 and the virtual reality display device 110 can communicate with each other through a network 112”). Allowable Subject Matter Claim 12 is objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims. The following is a statement of reasons for the indication of allowable subject matter: Soares (US 2026/0094354) discloses that Gaussians are based on level of details (Fig. 5). However, the limitations “generating a set of splats that form a third surface that matches the second surface formed by the set of 3D primitives by a different second acceptable amount of structural loss,” “defining visual characteristics for the set of splats to match the visual characteristics across the second surface formed by the set of 3D primitives by a different second acceptable amount of visual characteristic loss,” and “generating a first splat representation of the 3D object at a first fidelity with the one or more splats and a second splat representation of the 3D object at a second fidelity with the set of splats, wherein the second fidelity is greater than the first fidelity,” taken as a whole, render the claim patentably distinct over the prior art. Conclusion The prior art made of record and not relied upon is considered pertinent to applicant's disclosure: US 2024/0355047 – Gaussian splats US 2025/0157114 – Gaussian splats Any inquiry concerning this communication or earlier communications from the examiner should be directed to ANH-TUAN V NGUYEN whose telephone number is 571-270-7513. The examiner can normally be reached on M-F 9AM-5PM ET. 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, JASON CHAN can be reached on 571-272-3022. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300. Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000. /ANH-TUAN V NGUYEN/ Primary Examiner, Art Unit 2619
Read full office action

Prosecution Timeline

Jan 02, 2025
Application Filed
Sep 08, 2026
Non-Final Rejection mailed — §103 (current)

Precedent Cases

Applications granted by this same examiner with similar technology

Patent 12776095
VIDEO COMMENT DISPLAY METHOD AND APPARATUS, TERMINAL, AND STORAGE MEDIUM
2y 3m to grant Granted Oct 27, 2026
Patent 12749237
GENERATING AN INTERMEDIATE REPRESENTATION FOR CONVERTING BETWEEN LEGACY VECTOR PATHS AND VECTOR NETWORKS
2y 1m to grant Granted Sep 29, 2026
Patent 12657824
SYSTEMS AND METHODS FOR FACE ASSET CREATION AND MODELS FROM ONE OR MORE IMAGES
2y 8m to grant Granted Jun 16, 2026
Patent 12626456
ELECTRONIC DEVICE FOR DISPLAYING VIRTUAL OBJECT AND OPERATION METHOD THEREOF
2y 5m to grant Granted May 12, 2026
Patent 12614358
AUGMENTED REALITY ENVIRONMENT MELDING
3y 1m to grant Granted Apr 28, 2026
Study what changed to get past this examiner. Based on 5 most recent grants.

Strategy Recommendation AI-generated — please review before filing

Get a prosecution strategy drawn from examiner precedents, rejection analysis, and claim mapping.
Typically takes 5-10 seconds — AI-generated, attorney review required before filing

Prosecution Projections

1-2
Expected OA Rounds
72%
Grant Probability
91%
With Interview (+19.1%)
2y 10m (~1y 0m remaining)
Median Time to Grant
Low
PTA Risk
Based on 512 resolved cases by this examiner. Grant probability derived from career allowance rate.

Sign in with your work email

Enter your email to receive a magic link. No password needed.

Personal email addresses (Gmail, Yahoo, etc.) are not accepted.

Free tier: 3 strategy analyses per month