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 .
Election/Restrictions
Applicant's election with traverse of Invention I, Species I.A (claims 1–6) in the reply filed on 04/10/2026 is acknowledged. The traversal is on the ground(s) that the use of computer-assisted searching of multiple classifications and the fact that the inventions may encompass similar subject matter. This is not found persuasive because the groups and species are set forth to independent and distinct technical features such that the search is diverse for each group ( Invention I is directed to training an occupancy classifier using aggregated voxel features obtained by projecting world voxels to multi-frame camera images. Invention II is directed to runtime temporal aggregation with noise reduction to generate occupancy estimations at times t and t-1 using anchor voxels, coordinate transformations, and density estimation. Species I.A is directed to projecting a voxel grid or performing aggregation using an average, weighted average, or deformable attention. Species I.B is directed to training the occupancy classifier by comparing voxel grid features to a ground truth value.)
Thus, while the restricted inventions may be disclosed as usable together in the same temporal multi-frame occupancy estimation system, the groups are directed to different underlying concepts and different technical approaches: a machine-learning training methodology based on multi-frame voxel projection and feature aggregation, and a runtime perception pipeline based on anchor-voxel identification, point cloud coordinate transformation, and density-based noise reduction. A search directed to one group would not necessarily be expected to identify the most relevant prior art for the other group. For example, a search for multi-frame voxel feature aggregation and occupancy classifier training techniques (CPC classes G06V10/40, G06V10/82, G06V10/764) would not necessarily locate the most relevant prior art for runtime anchor-voxel noise reduction using point cloud transformation and density estimation; likewise, a search for temporal aggregation and noise reduction in vehicle perception systems (CPC classes G06V20/56, G06V20/58) would not necessarily locate the most relevant prior art for voxel feature aggregation methods and occupancy classifier training.
Accordingly, the search for the generic claims would not reasonably encompass the specific subject matter of each dependent subcombination or species, and examination of all groups together would impose a serious search and examination burden. For at least these reasons, and upon reconsideration of Applicant’s traversal, the restriction requirement is still deemed proper and is therefore made FINAL.
Claim Objections
Claim 1 is objected to because of the following informalities: The phrase “projecting each of a plurality of world voxels to the camera” is imprecise because the specification describes projecting voxel locations to an image plane or image coordinates associated with the camera, rather than physically projecting voxels “to the camera”. Appropriate correction would be to amend the phrase to “projecting each of a plurality of world voxels to an image plane of the camera” or “to image coordinates associated with the camera”. Appropriate correction is required.
Claim 4 is objected to because of the following informalities: Claim 4 recites “wherein projecting each of a plurality of world voxels to the camera at the first time t and the second time t−1 further comprises: performing voxel feature aggregation…” However, claim 1 separately recites projecting the plurality of world voxels and aggregating voxel features as distinct operations. Therefore, claim 4 appears to improperly characterize voxel feature aggregation as part of the projecting step, rather than as a further limitation of the aggregating step. Appropriate correction would be to amend claim 4 to recite, for example, “wherein aggregating voxel features comprises performing voxel feature aggregation…” or otherwise clarify the relationship between the projecting step and the voxel-feature aggregation step. Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 3 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 3 recites the limitation “extracting the feature from the second image captured at the second time t−1”. There is insufficient antecedent basis for this limitation in the claim. Specifically, Claim 2 recites "extracting a feature from the first image". Claim 3 recites "extracting the feature from the second image". It is unclear if "the feature" in Claim 3 refers back to the exact feature extracted from the first image in Claim 2 (which is technically impossible or illogical for a separate second image), or if it is intended to mean extracting a distinct, second feature from the second image. Applicant should amend Claim 3 to recite "extracting a feature from the second image" to resolve this ambiguity.
In the present instance, claim 3 further recites the broad recitation “projecting the voxel grid definition in the local coordinate system at the second time t−1 to the second image”. However, it is unclear whether “the voxel grid definition” refers to the voxel grid definition projected to the first image in claim 2, a transformed voxel grid definition corresponding to the second time t−1, or a separate voxel grid definition for the second image. The specification describes transforming the voxel grid definition to the local coordinate system of the vehicle at prior times, including t−1, before projecting the transformed voxel-grid definition onto the second image. As presently written, claim 3 does not clearly recite whether such transformation is required. Accordingly, the metes and bounds of claim 3 are unclear.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claim 1 is rejected under 35 U.S.C. § 101 because the claimed invention is directed to a judicial exception without significantly more.
The flow chart in MPEP § 2106, Subject Matter Eligibility Test for Products and Processes, is referred to for establishing ineligible subject matter. The USPTO's current subject matter eligibility guidance is found in MPEP §§ 2103–2106.07, and the Office's 2024 AI SME Update and August 4, 2025 memorandum further explain the analysis for computer-implemented and AI-related claims.
Step 1: Statutory Category
Independent claim 1 recites a "computer-implemented method," which falls within the process category of statutory subject matter. Therefore, claim 1 satisfies Step 1.
Step 2A, Prong One: Judicial Exception
Claim 1 recites the following steps: (1) Receiving a first image captured by a camera of a vehicle at a first time t; (2) Receiving a second image captured by the camera of the vehicle at a second time t-1; (3) Projecting each of a plurality of world voxels to the camera at the first time t and the second time t-1; (4) Aggregating voxel features for the plurality of world voxels for the first image and the second image; and (5) Training an occupancy classifier using the aggregated voxel features.
Steps (1) and (2) constitute mere data gathering, they recite only the passive receipt of image data captured by a camera. Such data collection steps are recognized as adding no meaningful weight to the eligibility analysis. See MPEP §2106.05(g) (noting that "insignificant extra-solution activity" such as mere data gathering does not integrate an abstract idea into a practical application). Steps (3), (4), and (5) collectively recite mathematical concepts: coordinate transformation (projecting world voxels to a camera view), a mathematical aggregation operation (aggregating voxel features), and a mathematical optimization/ training process (training an occupancy classifier). These steps fall squarely within the "mathematical concepts" category of abstract ideas enumerated in the 2019 Revised Guidance, as they describe mathematical operations performed on numerical data representing spatial coordinates and learned model parameters. See 2019 Revised Guidance, 84 Fed. Reg. 50 (Jan. 7, 2019); see also 2024 USPTO Guidance Update on Patent Subject Matter Eligibility, 89 Fed. Reg. 58128 (July 17, 2024).
Accordingly, Claim 1 recites a judicial exception: specifically, mathematical concepts including mathematical relationships and mathematical operations.
Step 2A, Prong Two: No Integration into a Practical Application
The examiner has considered whether Claim 1 as a whole integrates the recited abstract idea into a practical application such that the claim is directed to something more than the abstract idea itself. For the reasons below, it does not.
The vehicle camera limitation is a mere field-of-use designation. The recitation that the images are "captured by a camera of a vehicle" identifies only the environment in which data is collected. This does not impose any meaningful structural or functional constraint on the claimed method beyond specifying the source of input data. A field-of-use limitation that merely identifies where or how data originates does not constitute integration into a practical application. See MPEP §2106.05(h); Intellectual Ventures I LLC v. Capital One Bank, 792 F.3d 1363, 1366 (Fed. Cir. 2015) (field-of-use limitations do not make an otherwise abstract claim patent eligible).
The claim does not recite a specific technical improvement to computer functionality or a technical field. Claim 1 recites only that voxels are projected and features are aggregated at a high level of generality. It does not specify how the voxel projection is performed, what coordinate system or transformation is used, how features are aggregated, or what architecture the occupancy classifier uses. The claim is therefore distinguishable from claims that specify a particular technical mechanism that improves computer processing, image analysis, or perception system performance. The Federal Circuit in Recentive Analytics v. ESPN (Fed. Cir. 2025) emphasized that claims applying generic machine learning to a new domain, without specifying technical improvements to the underlying ML process, are directed to an abstract idea without a practical application.
The claim recites a result, not a technical solution. The final step "training an occupancy classifier using the aggregated voxel features" recites a desired outcome (a trained classifier) without specifying any concrete technical means for achieving it. Result-oriented claiming without a disclosed concrete implementation does not provide the specificity necessary to satisfy Prong 2. See MPEP §2106.05(f); Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1356 (Fed. Cir. 2016) (claims focused on "the combination of those results" without specifying how results are achieved are directed to an abstract idea).
Regarding dependent claims: the examiner notes that dependent claims 2–6 recite additional technical details, including extracting features, projecting a voxel grid in a local coordinate system, performing feature sampling, voxel feature aggregation, and the use of deformable attention, that may supply the specific technical hooks needed for a practical application. However, Claim 1 as drafted does not incorporate these limitations and is evaluated based on its own recited elements. At the breadth of Claim 1, the recited method covers any and all ways of projecting voxels and aggregating features to train any occupancy classifier, an overbreadth that forecloses a finding of practical application integration.
Accordingly, Claim 1 does not integrate the recited mathematical concepts into a practical application and fails Step 2A, Prong 2.
Step 2B: No Significantly More / No Inventive Concept
Because Claim 1 fails Step 2A, Prong 2, the examiner proceeds to Step 2B. The examiner has considered whether Claim 1 recites additional elements, individually or in combination, that amount to significantly more than the recited abstract idea.
The additional elements recited in Claim 1 beyond the abstract idea are: a camera of a vehicle (data-gathering hardware); and an implied computer executing the method (by virtue of the preamble "computer implemented method").
Both elements represent generic computer components performing generic computer functions, receiving data and executing mathematical operations at a level well understood, routine, and conventional in the field of autonomous vehicle perception and computer vision at the time of filing. See Alice Corp. v. CLS Bank Int'l, 573 U.S. 208, 225–26 (2014) (generic computer implementation insufficient to transform an abstract idea into patent-eligible subject matter). The mere invocation of a computer and a camera does not supply an inventive concept.
Accordingly, Claim 1 does not recite significantly more than the abstract idea itself, and the rejection under §101 is proper. Applicant is invited to amend Claim 1 to incorporate the technical specificity present in the dependent claims, to strengthen the practical application.
Claim Rejections - 35 USC § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1–6 are rejected under 35 U.S.C. §102(a)(1) as being anticipated by Wang (Wang et al, “PANOOCC: Unified Occupancy Representation for Camera-based 3D panoptic segmentation”. arXiv.org., 2023).
Regarding claim 1, Wang teaches a computer-implemented method comprising:
receiving a first image captured by a camera of a vehicle at a first time t;
receiving a second image captured by the camera of the vehicle at a second time t-1;
( [Sec. 3.1-3.2], [Fig. 2]: Wang teaches taking current multi-view images
I
t
=
{
I
t
1
,
I
t
2
,
.
.
.
,
I
t
n
}
and previous frames
I
t
-
1
,
.
.
.
,
I
t
-
k
as input, where n denotes a camera-view index and k denotes the number of history frames; wherein the current multi-view images
I
t
correspond to the first image captured at the first time t, and the previous frames
I
t
-
1
correspond to the second image captured at the second time t−1, both captured by cameras of the ego-vehicle in the autonomous driving context. Fig. 2 depicts current frame T and previous frames including T−1. Wang also teaches use of the nuScenes autonomous-driving dataset, where each sample includes RGB images from six cameras. )
projecting each of a plurality of world voxels to the camera at the first time t and the second time t-1;
( [Sec. 3.3-3.4], [Eq. 1–3], [Eq. 6–7]: Wang teaches defining 3D-grid-shape voxel queries Q ∈
R
H
×
W
×
Z
×
D
, where each voxel query at position (i, j, k) is responsible for a corresponding 3D voxel-grid cell region, and each grid cell corresponds to a real-world size. Wang further teaches voxel cross-attention in which πn(Refm i,j,k) denotes a projected reference point in the n-th camera view, projected by projection matrix πn from the voxel grid located at (i, j, k). Wang also teaches the projection between a real 3D reference point in the ego-vehicle frame and a corresponding 2D reference point on the n-th image view using projection matrix Pn. Wang further teaches temporal voxel alignment between the current frame and previous frames using transformation matrix
T
t
→
t
-
k
, including previous frame t−1. )
aggregating voxel features for the plurality of world voxels for the first image and the second image; and
( [Sec. 3.2], [Sec. 3.4], [Fig. 2]: Wang teaches using voxel queries to aggregate spatiotemporal information. Wang further teaches that an image backbone extracts multi-scale features, a View Encoder uses voxel queries to learn voxel features, and a Temporal Encoder aligns and fuses previous voxel features with the current frame. Wang also teaches that previous aligned voxel queries [world voxels for the second image at time t-1] are concatenated with current voxel queries [world voxels for the first image at time t] and fused to output fused-voxel queries Qf. )
training an occupancy classifier using the aggregated voxel features.
( [Sec. 3.1], [Sec. 3.5-3.6], [Sec. 4.2], [Eq. 8–9]: Wang teaches camera-based 3D semantic occupancy prediction for predicting semantic voxel volume and determining whether a voxel grid is empty or occupied. Wang further teaches converting fused voxel features Qf into occupancy features O, training the model end-to-end based on the unified occupancy representation, and using a lightweight MLP segmentation head based on occupancy features to query voxel-grid status at arbitrary positions. Wang also teaches supervising voxel prediction using segmentation losses during training. )
Regarding claim 2, Wang teaches the computer-implemented method of claim 1, wherein projecting each of a plurality of world voxels to the camera at the first time t and the second time t-1 comprises:
extracting a feature from the first image captured at the first time t;
( [Sec. 3.2], [Sec. 3.4], [Fig. 2], [Sec. 4.2]: PanoOcc takes multi-frame multi-view images as input and first extracts multi-scale features using an image backbone; wherein the image backbone (ResNet50, ResNet101-DCN, or InternImage) extracts multi-scale image features
F
from the current multi-view images
I
t
at the first time t. )
projecting a voxel grid definition in a local coordinate system at the first time t to the first image; and
performing feature sampling for the first image based at least in part on results of the extracting and results of the projecting.
( [Sec. 3.3-3.4], [Eq. 1–3], [Fig. 2]: Wang teaches defining 3D-grid-shape voxel queries Q ∈
R
H
×
W
×
Z
×
D
, wherein the voxel queries are positioned at grid coordinates (i, j, k), the local coordinate system at the first time t. Wang further discloses the projection of this voxel grid definition to the first image via Eq. 2–3, where
P
n
is the projection matrix of the n-th camera. Feature sampling is then performed at the resulting projected 2D reference points using a voxel cross-attention mechanism via Eq. 1, using the projected reference point in the n-th camera view projected by projection matrix
π
n
from the voxel grid located at (i,j,k) and the extracted image features Fn. )
Regarding claim 3, Wang teaches the computer-implemented method of claim 2, wherein projecting each of a plurality of world voxels to the camera at the first time t and the second time t-1 further comprises:
extracting the feature from the second image captured at the second time t-1;
( [Sec. 3.2], [Sec. 3.4], [Fig. 2], [Sec. 4.2]: the same image backbone discussed in claim 2 processes all temporal frames including the previous frames
I
t
-
1
,
.
.
.
,
I
t
-
k
at the second time t−1, extracting multi-scale features
F
from the second image at time t−1 in the same manner. Fig. 2 depicts both current frame T and previous frames including T−1 being processed. )
projecting the voxel grid definition in the local coordinate system at the second time t-1 to the second image; and
performing feature sampling for the second image based at least in part on results of the extracting and results of the projecting.
( [Sec. 3.3-3.4], [Eq. 1–3], [Eq. 6–7]: Eq. 6 explicitly transforms the voxel grid definition
G
t
, defined in the local coordinate system at the first time t, into
G
t
-
1
, which is the voxel grid definition expressed in the local coordinate system at the second time t-1. Feature sampling for the second image is then performed via Eq. 7, wherein GridSample performs interpolation sampling of the prior-frame voxel queries
Q
t
-
1
. Wang further discloses: "the queries at frame t−k are aligned to current frame t by interpolation sampling, denoted as
Q
t
-
k
→
t
" thereby performing feature sampling for the second image based at least in part on both the results of the extracting of features from the second image at t-1 and the results of the projecting of the voxel grid definition into the local coordinate system at t-1. )
Regarding claim 4, Wang teaches the computer-implemented method of claim 3, wherein projecting each of a plurality of world voxels to the camera at the first time t and the second time t-1 further comprises:
performing voxel feature aggregation based at least in part on results of the feature sampling for the first image and results of the feature sampling for the second image; and
( [Sec. 3.2], [Sec. 3.4-3.5], [Fig. 2]: Wang teaches using voxel queries to aggregate spatiotemporal information from multi-frame and multi-view images. Wang specifically teaches after aligning previous voxel queries to the current frame, concatenating previous aligned voxel queries with current voxel queries, and using residual 3D convolution to fuse the queries and output fused voxel queries Qf; wherein the fused voxel queries
Q
f
∈
R
H
×
W
×
Z
×
D
represent voxel feature aggregation performed based on the feature sampling results for the first image at time t (
Q
t
from VCA, Eq. 1) and the feature sampling results for the second image at time t−1 (
Q
t
-
1
→
t
from GridSample, Eq. 7). )
wherein training the occupancy classifier comprises generating an occupancy estimation network based at least in part on the voxel feature aggregation.
( [Sec. 3.2], [Sec. 3.5-3.6], [Fig. 2]: Wang teaches an Occupancy Encoder that outputs fused voxel features Qf, an Occupancy Decoder that converts the fused voxel feature Qf into fine-grained occupancy features O, and a Task Head including a segmentation head. Wang further teaches that the model is trained end-to-end based on the unified occupancy representation, and that the segmentation head is a lightweight MLP based on occupancy features O or Osparse that queries voxel-grid status at arbitrary positions. )
Regarding claim 5, Wang teaches the computer-implemented method of claim 4, wherein the occupancy estimation network is generated using voxel grid features.
( [Sec. 3.3-3.6], [Sec. 4.1]: Wang teaches voxel queries arranged in a 3D grid shape
Q
∈
R
H
×
W
×
Z
×
D
, , with each voxel query corresponding to a 3D voxel-grid cell region. Wang further teaches that, given voxel queries Q and extracted image features F, the Occupancy Encoder outputs fused voxel features
Q
f
∈
R
H
×
W
×
Z
×
D
. Wang then teaches converting the fused voxel feature Qf into occupancy features O, and using an MLP segmentation head based on occupancy features to query voxel-grid status [corresponding to generating the occupancy-estimation network using voxel-grid features]. )
Regarding claim 6, Wang teaches the computer-implemented method of claim 4, wherein the voxel feature aggregation is performed using at least one of an average, a weighted average, or a deformable attention.
( [Sec. 3.4], [Eq. 1], [Table 8]: Wang explicitly states "we draw inspiration from the querying paradigm in recent BEV-based methods and adopt efficient deformable attention for voxel cross-attention and voxel self-attention", directly naming deformable attention as the aggregation mechanism. Wang’s voxel cross-attention is formulated using
D
A
(
q
,
π
n
(
R
e
f
i
,
j
,
k
m
)
,
F
n
)
, where DA represents deformable attention applied between voxel queries, projected reference points, and image features. Wang further confirms at Table 8 ablation: "DA means deformable attention" thereby teaching voxel feature aggregation performed using at least one of an average, a weighted average, or a deformable attention.)
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KEN KUDO whose telephone number is (571)272-4498. The examiner can normally be reached M-F 8am - 5pm.
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, Vincent Rudolph can be reached at 571-272-8243. 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.
KEN KUDO
Examiner
Art Unit 2671
/KEN KUDO/Examiner, Art Unit 2671
/VINCENT RUDOLPH/Supervisory Patent Examiner, Art Unit 2671
Prosecution Insights