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 the Applicant’s arguments
The previous rejection is withdrawn. Applicant’s amendments are entered. Applicant’s remarks are also entered into the record. A new search was made necessitated by the applicant’s amendments.
A new reference was found. A new rejection is made herein.
Applicant’s arguments are now moot in view of the new rejection of the claims.
PNG
media_image1.png
748
706
media_image1.png
Greyscale
Claim 1 is amended to recite and the primary reference is silent but RANKAWAT et al. teaches “...training a machine-learning model to classify underdrivable regions and non- underdrivable regions based on the ground truth data; and
autonomously driving a vehicle using the trained machine-learning model”. (see paragraph 36-38 122-127 where the DNN can use the ground truth data for predictions. See FIG. 6 where the ground truth data can predict the traversable boundary and the non-traversable boundary for the autonomous vehicle and then the vehicle can be controlled to drive on the road free of obstacles)”.
It would have been obvious for one of ordinary skill in the art before the effective filing date to combine the disclosure of LADD HA with the teachings of RANKAWAT with a reasonable expectation of success since RANKAWAT teaches that a ground truth data can be determined from the LIDAR data sensor data. Ground truth data is verified, accurate, real-world data used to train and validate artificial intelligence (AI) and machine learning models. It serves as correct answer against which a model's predictions are compared. Without high-quality ground truth data, models may learn incorrect patterns, leading to flawed or unreliable results. In this way, a boundary and a non-boundary can be determined in FIG. 6. See paragraph 1-8 of RANKAWAT of NVIDIA™.
Notice of Pre-A/A or AJA Status
The present application, filed on or after March 16, 2013, is being examined under the first
inventor to file provisions of the AIA .
Drawings
The drawings were received on June 17th 2022. These drawings are accepted.
Priority
Acknowledgment is made of applicant's claim for foreign priority based on an application in
specification filed on July 28th 2022.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-3, 5-8,12,13 and 16 and 18 rejected under 35 U.S.C. 103 as being unpatentable as obvious in view of Ladd ha {Patent No. US11762094B2) and in view of German Patent Pub. No.: DE112021000422T5 (US20210295171A1) to Kamenev et al. that was filed in 2020 (hereinafter “KAMENEV”) and in view of United States Patent Application Pub. No.: US 2019/02861531 Rankawat et al. that was filed in 2019 (hereinafter “Rankawat”).
Regarding claim 1 Ladd ha teaches a computer-implemented method for generating ground
truth data, the method comprising; (See Ladd ha column 1, line40-44; "disclosure is directed to a
computer-implemented method. The method can include obtaining, by a computing system with one or
more processors, a plurality of sensor sweeps from a sensor associated with an autonomous vehicle";
Also see Laddha column 5, line45-58; "vehicle's surrounding environment, the vehicle computing
system can access sensor data from one or more sensors (e.g., LI DAR, RADAR, camera, etc.) to identify
static objects and/or dynamic objects (actors) in the autonomous vehicle's environment. To help
determine its position within the environment (and relative to these objects), the vehicle computing
system can provide sensor data to a machine-learned model(s ). In addition, or alternatively, the
autonomous vehicle can access map data (e.g., high definition map data, etc.) to determine the
autonomous vehicle's current position relative to other objects in the world (e.g., bicycles, pedestrians,
other vehicles, buildings, etc.), as well as map features such as, for example, lane boundaries, curbs, and
so on." (Examiner notes- ground truth data - vehicle position with in environment, object identification
and position relative to objects)); for a plurality of points in time, acquiring sensor data for a respective
point in time; and for at least a subset of the plurality of points in time, determining ground truth data
of the respective point in time based on the sensor data of a future point of time and at least one of a
present point of time or a past point of time; (See Ladd ha column 1, line 40-44; "disclosure is directed
to a computer-implemented method. The method can include obtaining, by a computing system with
one or more processors, a plurality of sensor sweeps from a sensor associated with an autonomous
vehicle";
Also Ladd ha column 1-2, line 64-4; "The method can include mapping, by the computing
system, each point from the respective image to the image associated with the next time step to
generate a fused image. The method can include generating, by the computing system, a final fused
representation of the plurality of sensors sweeps once a sensor sweep associated with the current time
step has been combined with all previously generated sensor sweeps.").
Claim 1 is amended to recite and LADD HA is silent but Kamenev teaches “...the sensor data of the future point of time being based on sequence of historic sensor data”. (see paragraph 60-79 where sensor data can be mined to build a DNN data set and provide a confidence map of future trajectories of the vehicle and actors with a confidence level. The autonomous vehicle can use both the 1. Past and 2. Future trajectories from stereo vision camera and audio sensors; and IMU, LIDAR and RADAR sensors in paragraph 128).
PNG
media_image2.png
806
544
media_image2.png
Greyscale
It would have been obvious for one of ordinary skill in the art before the effective filing date of the present disclosure to combine the disclosure of LADD HA with the teachings of KAMANEV with a reasonable expectation of success since KAMANEV teaches that a DNN can provide a number of future trajectories based on the past trajectories and the second data. As such, the DNN 116 may learn to predict future trajectories—or information representative thereof—by monitoring and factoring in past locations of actors, road structures, wait conditions, and/or other information over a plurality of time slices. In some embodiments, the DNN 116 may include a recurrent neural network (RNN). For a non-limiting example, and as described in more detail below with respect to FIG. 1B, the DNN 116 may include an encoder-decoder RNN 116A. This can provide past and future locations of the actor for collision avoidance. See claims 1-4 and paragraph 34-44.
PNG
media_image1.png
748
706
media_image1.png
Greyscale
Claim 1 is amended to recite and the primary reference is silent but RANKAWAT et al. teaches “...training a machine-learning model to classify underdrivable regions and non- underdrivable regions based on the ground truth data; and
autonomously driving a vehicle using the trained machine-learning model”. (see paragraph 36-38 122-127 where the DNN can use the ground truth data for predictions. See FIG. 6 where the ground truth data can predict the traversable boundary and the non-traversable boundary for the autonomous vehicle and then the vehicle can be controlled to drive on the road free of obstacles)”.
It would have been obvious for one of ordinary skill in the art before the effective filing date to combine the disclosure of LADD HA with the teachings of RANKAWAT with a reasonable expectation of success since RANKAWAT teaches that a ground truth data can be determined from the LIDAR data sensor data. Ground truth data is verified, accurate, real-world data used to train and validate artificial intelligence (AI) and machine learning models. It serves as correct answer against which a model's predictions are compared. Without high-quality ground truth data, models may learn incorrect patterns, leading to flawed or unreliable results. In this way, a boundary and a non-boundary can be determined in FIG. 6. See paragraph 1-8 of RANKAWAT of NVIDIA™.
Regarding claim 2, Laddha teaches the computer-implemented method of claim 1, Ladd ha
further teaches wherein: at least one of the present point of time, the past point of time, or the future
point of time are relative to the respective point in time; (See Ladd ha column 4, line 1-11; "Thus, each
sweep represents a different time step between a point in the past and the most recent sweep
(effectively the current time). In addition, because the sensor sweeps a re captured by a LI DAR sensor
attached to the autonomous vehicle, as the autonomous vehicle moves, the coordinate frame of the
sensor data included in each sensor sweep changes. For example, a stationary mailbox may appear in
front of the autonomous vehicle in a sensor sweep associated a current time but beside or
behind the autonomous vehicle in a future sensor sweep as the autonomous vehicle moves past the
mailbox.").
Regarding claim 3, Laddha teaches the computer-implemented method of claim 1, Ladd ha
further teaches wherein: the sensor data includes at least one of radar data or lidar data; (See Ladd ha
column 5, line 46-48; "the vehicle computing system can access sensor data from one or more sensors
(e.g., LIDAR, RADAR, camera, etc.)").
Claim 4 is cancelled.
Regarding claim 5, Laddha teaches the computer-implemented method of claim 4, Ladd ha also
teaches wherein the machine-learning model is configured to at least one of: determine an occupancy
grid; or classify an object with respect to under drivability; (See Ladd ha column 8, line 37-43; "The
machine-learned model can use feature data to identify one or more objects within the fused image and
the past movement of those objects between the earliest sensor sweep and the most recent sensor
sweep. A prediction system can use this data to predict the future movement and trajectories of each of
the one or more identified objects.").
Regarding claim 6, Laddha teaches the computer-implemented method of claim 5, Ladd ha also
teaches wherein the determining comprises: determining the ground truth data based on at least two
maps; (See Ladd ha column 5, line53-58; "the autonomous vehicle can access map data (e.g., high
definition map data, etc.) to determine the autonomous vehicle's current position relative to other
objects in the world (e.g., bicycles, pedestrians, other vehicles, buildings, etc.), as well as map features
such as, for example, lane boundaries, curbs, and so on.").
Regarding claim 7, Laddha teaches the computer-implemented method of claim 6, Ladd ha also
teaches wherein: the at least two maps include a full-range map based on scans that are irrespective
of a range of the scans; (See Ladd ha column 17, line 23-34; "the vehicle computing system 110 in
processing, analyzing, and perceiving its surrounding environment and its relations hip thereto. In some
implementations, the map data 160 can include high definition map data. In some implementations, the
map data 160 can include sparse map data indicative of a limited number of environmental features
(e.g., lane boundaries, etc.). In some implementations, the map data can be limited to geographic
area(s) and/or operating domains in which the vehicle 105 (or autonomous vehicles generally) may
travel (e.g., due to legal/regulatory constraints, autonomy capabilities, and/or other factors)."; Also see
Ladd ha column 4, line 15-21; "a vehicle computing system associated with the autonomous vehicle can
let the data remain in a range view. To be able to use the sensor data for detection and prediction, the
vehicle computing system can convert each sensor sweep (e.g., the data gathered in a full revolution of
the LIDAR sensor)into at wo-dimensional image in range view.").
Regarding claim 8, Ladd ha teaches the computer-implemented method of claim 7, Ladd ha also
teaches wherein: the at least two maps include a limited-range map based on scans that are below a
pre-determined range threshold; (See Ladd ha column 17, line23-34; "the vehicle computing system
110 in processing, analyzing, and perceiving its surrounding environment and its relationship thereto. In
some implementations, the map data 160 can include high definition map data. In some
implementations, the map data 160 can include sparse map data indicative of a limited number of
environmental features (e.g., lane boundaries, etc.). In some implementations, the map data can be
limited to geographic area(s) and/or operating domains in which the vehicle 105 (or autonomous
vehicles generally) may travel (e.g., due to legal/regulatory constraints, autonomy capabilities, and/or
other factors)."; Also see Ladd ha column 4, line 15-21; "a vehicle computing system associated with the
autonomous vehicle can let the data remain in a range view. To be able to use the sensor data for
detection and prediction, the vehicle computing system can convert each sensor sweep (e.g., the data
gathered in a full revolution of the LI DAR sensor) into a two-dimensional image in range view.").
With respect to independent claims 12 and 16, please see the rejection above with respect to
claim 1 which is commensurate in scope to claims 12 and 16, with claim 1 being drown computer implemented method, and claim 12 being drawn to a corresponding non-transitory computer-readable
medium, claim 18 being drawn to a corresponding system.
PNG
media_image1.png
748
706
media_image1.png
Greyscale
Claims 12-16 are amended to recite and the primary reference is silent but RANKAWAT et al. teaches “...training a machine-learning model to classify underdrivable regions and non- underdrivable regions based on the ground truth data; and
autonomously driving a vehicle using the trained machine-learning model”. (see paragraph 36-38 122-127 where the DNN can use the ground truth data for predictions. See FIG. 6 where the ground truth data can predict the traversable boundary and the non-traversable boundary for the autonomous vehicle and then the vehicle can be controlled to drive on the road free of obstacles)”.
It would have been obvious for one of ordinary skill in the art before the effective filing date to combine the disclosure of LADD HA with the teachings of RANKAWAT with a reasonable expectation of success since RANKAWAT teaches that a ground truth data can be determined from the LIDAR data sensor data. Ground truth data is verified, accurate, real-world data used to train and validate artificial intelligence (AI) and machine learning models. It serves as correct answer against which a model's predictions are compared. Without high-quality ground truth data, models may learn incorrect patterns, leading to flawed or unreliable results. In this way, a boundary and a non-boundary can be determined in FIG. 6. See paragraph 1-8 of RANKAWAT of NVIDIA™.
With respect to dependent claim 13, please see the rejection above with respect to
combination of claims 4 and 5 which are commensurate in scope to claim 13, with combination of claims
4 and 5 being drown computer-implemented method, and claim 13 being drawn to a corresponding
non-transitory computer-readable medium.
Claim 17 is cancelled.
Regarding claim 18, Laddha teaches the system of claim 17, Ladd ha also teaches, wherein
the machine-learning model comprises an artificial neural network; (See Ladd ha column 4, line 26-37;
"vehicle computing system can determine the motion of the autonomous vehicle during a particular
period. For example, the vehicle computing system can access a location determination system (e.g., a
GPS system) and determine the current position of the autonomous vehicle for each sensor sweep. The
vehicle computing system can, for each sweep in the plurality of sweeps using a machine-learned model
such as a convolutional neural network, extract feature data representing one or more features in the
image. For example, features can include low-level components of an image including shapes, edges,
lines, blobs, points, and soon.").
Claims 9-11,14,15,19 and 20 rejected under 35 U.S.C. 103 as being unpatentable over Ladd ha
{Patent No. US11762094B2) in view of Smolyanskiy {Patent No. US20220415059Al) and in view of KAMENEV and Rankawat.
Regarding claim 9, Laddha teaches the computer-implemented method of claim 8, Ladd ha does
not teach but Smolyanskiy teaches further comprising: labeling a cell as non-under drivable or
under drivable based on a probability of the cell in the full-range map and a probability of the cell in
the limited-range map; (See Smolyanskiy paragraph0037; " ... LiDAR range image ... outputs may be
processed into 2D and/or 3D bounding boxes and class labels for the detected objects. In an example
application for autonomous vehicles, the DNN may be used to predict one or more bounding boxes (e.g.,
2D bounding box in top-down view, 3D bounding box) for each detected object on the road or sidewalk,
a class label for each detected object, and a 2D mask demarcating a drivable space, sidewalks, buildings,
trees, poles, other static environmental parts (e.g., in the top-down view). In some embodiments, 2D
bounding boxes in top-down view may be adapted into 3D bounding boxes by deriving box height from
the predicted object data."; Also see Smolyskiy paragraph 0061; " ... confidence map for a
particular class as an example, the confidence map may have spatial dimensions corresponding to an
input into the encoder/decoder 605 (e.g., a LiDAR range image), and the confidence map may include a
classification value for each pixel (e.g., a probability, score, or log it). In some cases, the classification
values may be mapped to known 3D locations identified by corresponding sensor data 402 and/or input
data 406.").
Both Laddha and Smolyanskiy are in the same field of endeavor of method and systems for
generating ground truth data. It would have been obvious for one ordinary skilled in the art before the
effective filing date of present invention to modify Laddha systems and methods for object detection
with the primary reference a be ling a cell based on a probability of the cell in the full-range map and a probability of the cell in the limited-range map. No new functionality would arise from the combination and the combination would improve usability of Laddha by adding the labeling a cell based on a probability of the cell in the full-range map and a probability of the cell in the limited-range map, one of ordinary skill in the art would have recognized that the results of the combination were predictable.
Regarding claim 10, Laddha teaches the computer-implemented method of claim 9, Laddha
does not teach but Smolyanskiy teaches wherein the labeling comprises: labeling the cell as non underdrivable responsive to the probability of the cell in the limited-range map being above a first
pre-determined threshold;(See Smolyanskiy paragraph 0037; " ... LiDAR range image ... outputs may be
processed into 2D and/or 3D bounding boxes and class labels for the detected objects. In an example
application for autonomous vehicles, the DNN may be used to predict one or more bounding boxes (e.g.,
2D bounding box in top-down view, 3D bounding box) for each detected object on the road or sidewalk,
a class label for each detected object, and a 2D mask demarcating a drivable space, sidewalks, buildings,
trees, poles, other static environmental parts (e.g., in the top-down view). In some embodiments, 2D
bounding boxes in top-down view may be adapted into 3D bounding boxes by deriving box height from
the predicted object instance data."; Also see Smolyanskiy paragraph 0061; " ... confidence map for a
particular class as an example, the confidence map may have spatial dimensions corresponding to an
input into the encoder/decoder 605 (e.g., a LiDAR range image), and the confidence map may include a
classification value for each pixel (e.g., a probability, score, or log it). In some cases, the classification
values may be mapped to known 3D locations identified by corresponding sensor data 402 and/or input
data 406."; further see Smolyanskiy paragraph0076; " ... candidate bounding shapes that have a
confidence/probability of being a member of the object class less than some threshold (e.g., 50%) may
be filtered out. Additionally or alternatively, a candidate bounding box (or other shape) with the highest
confidence/probability score for a particular class may be assigned an instance ID, a metric such as
intersection over union (loU) may be calculated with respect to each of the other candidates in the class,
and candidates having an loU above some threshold may be filtered out to remove duplicates.").
Both Ladd ha and Smolyanskiy are in the same field of endeavor of method and systems for
generating ground truth data. It would have been obvious for one ordinary skilled in the a rt before the
effective filing date of present invention to modify Laddha systems and methods for object detection
with Smolyanskiy labeling a cell based on a probability of the cell in the full-range map and a probability
of the cell in the limited-range map. No new functionality would arise from the combination and the
combination would improve usability of Laddha by adding the labeling a cell based on a probability of
the cell in the full-range map and a probability of the cell in the limited-range map, one of ordinary skill
in the art would have recognized that the results of the combination were predictable.
Regarding claim 11, Laddha teaches the computer-implemented method of claim 10, Laddha
does not teach but Smolyanskiy teaches wherein the labeling further comprises: labeling the cell as
underdrivable responsive to the probability of the cell in the full- range map being above a second
pre-determined threshold and the probability of the cell in the limited-range map being equal to a
value representing no occupation in the cell; (See Smolyanskiy paragraph0037; " ... LiDAR range image
... outputs may be processed into 2D and/or 3D bounding boxes and class labels for the detected
objects. In an example application for autonomous vehicles, the DNN may be used to predict one or
more bounding boxes (e.g., 2D bounding box in top-down view, 3D bounding box) for each detected
object on the road or sidewalk, a class label for each detected object, and a 2D mask demarcating a
drivable space, sidewalks, buildings, trees, poles, other static environmental parts (e.g., in the top-down
view). In some embodiments, 2D bounding boxes in top-down view may be adapted into 3D bounding
boxes by deriving box height from the predicted object instance data."; Also see Smolyanskiy paragraph
0061; " ... confidence map for a particular class as an example, the confidence map may have spatial
dimensions corresponding to an input into the encoder/decoder 605 (e.g., a LiDAR range image), and
the confidence map may include a classification value for each pixel (e.g., a probability, score, or log it).
In some cases, the classification values may be mapped to known 3D locations identified by
corresponding sensor data 402 and/or input data 406."; further see Smolyanskiy paragraph 0076;
" ... candidate bounding shapes that have a confidence/probability of being a member of the object class
less than some threshold (e.g., 50%) may be filtered out. Additionally or alternatively, a candidate
bounding box (or other shape) with the highest confidence/probability score for a particular class may
be assigned an ins ta nee ID, a metric such as intersection over union (loU) may be calculated with respect to each of the other candidates in the class, and candidates having an loU above some threshold may be filtered out to remove duplicates.").
Both Ladd ha and Smolyanskiy are in the same field of endeavor of method and systems for
generating ground truth data. It would have been obvious for one ordinary skilled in the a rt before the
effective filing date of present invention to modify Laddha systems and methods for object detection
with the primary reference with a cell based on a probability of the cell in the full-range map and a probability of the cell in the limited-range map. No new functionality would arise from the combination and the combination would improve usability of Ladd ha by adding the labeling a cell based on a probability of the cell in the full-range map and a probability of the cell in the limited-range map, one of ordinary skill in the art would have recognized that the results of the combination were predictable.
Regarding claim 14 Laddha teaches the non-transitory computer-readable medium of claim 12,
Ladd ha further teaches wherein the operations further include: training a machine-learning model
based on the ground truth data;(See Ladd ha column 12, line 36-58; "The means can be configured to
transform, using a machine learned model with the feature data and the movement data of the
autonomous vehicle as input, the feature data into a coordinate frame associated with a next time step.
For example, the vehicle computing system can determine where the features of the respective image
would have appeared had the LI DAR system been positioned at the same place as the next time step
when the measurement was taken. A transformation unit is one example of a means for transforming,
using a machine learned model with the feature data and the movement data of the autonomous
vehicle as input, the feature data into a coordinate frame associated with a next time step.").
Ladd ha does not teach but Smolyanskiy teaches the machine-learning model configured to
classify at least one of an object or a cell with respect to underdrivability or non-underdrivability; (See
Smolyanskiy paragraph0037; " ... LiDAR range image ... outputs may be processed into 2D and/or 3D
bounding boxes and class labels for the detected objects. In an example application for autonomous
vehicles, the DNN may be used to predict one or more bounding boxes (e.g., 2D bounding box in top down view, 3D bounding box) for each detected object on the road or sidewalk, a class label for each
detected object, and a 2D mask demarcating a drivable space, sidewalks, buildings, trees, poles, other
static environmental parts (e.g., in the top-down view). In some embodiments, 2D bounding boxes in
top-down view may be adapted into 3D bounding boxes by deriving box height from the predicted
object instance data.").
Both Ladd ha and Smolyanskiy are in the same field of endeavor of method and systems for
generating ground truth data. It would have been obvious for one ordinary skilled in the a rt before the
effective filing date of present invention to modify Ladd ha systems and methods for object detection
with Smolyanskiy labeling a cell based on a probability of the cell in the full-range map and a probability
of the cell in the limited-range map. No new functionality would arise from the combination and the
combination would improve usability of Ladd ha by adding the labeling a cell based on a probability of
the cell in the full-range map and a probability of the cell in the limited-range map, one of ordinary skill
in the art would have recognized that the results of the combination were predictable.
Regarding claim 15 Laddha teaches the non-transitory computer-readable medium of claim 12,
Ladd ha further teaches wherein the determining comprises: determining the ground truth data based
on at least two maps; (See Ladd ha column 5, line 53-58; "the autonomous vehicle can access map data
(e.g., high definition map data, etc.) to determine the autonomous vehicle's current position relative to
other objects in the world (e.g., bicycles, pedestrians, other vehicles, buildings, etc.), as well as map
features such as, for example, lane boundaries, curbs, and so on.").
Ladd ha does not teach but Smolyanskiy teaches the at least two maps including a full-range
map based on scans that are irrespective of a range of the scans and a limited-range map based on
scans that are below a pre-determined range threshold; (See Smolyanskiy paragraph0037; " ... LiDAR
range image ... outputs may be processed into 2D and/or 3D bounding boxes and class labels for the
detected objects. In an example application for autonomous vehicles, the DNN may be used to predict
one or more bounding boxes (e.g., 2D bounding box in top-down view, 3D bounding box) for each
detected object on the road or sidewalk, a class label for each detected object, and a 2D mask
demarcating a drivable space, sidewalks, buildings, trees, poles, other static environmental parts (e.g., in
the top-down view). In some embodiments, 2D bounding boxes in top-down view may be adapted into
3D bounding boxes by deriving box height from the predicted object instance data."; Also see
Smolyanskiy paragraph 0061; " ... confidence map for a particular class as an example, the confidence
map may have spatial dimensions corresponding to an input into the encoder /decoder 605 (e.g., a LiDAR range image), and the confidence map may include a classification value for each pixel (e.g., a
probability, score, or log it). In some cases, the classification values may be mapped to known 3D
locations identified by corresponding sensor data 402 and/or input data 406."; further see Smolyanskiy
paragraph 0076; " ... candidate bounding shapes that have a confidence/probability of being a member
of the object class less than some threshold (e.g., 50%) may be filtered out. Additionally or alternatively,
a candidate bounding box (or other shape) with the highest confidence/probability score for a particular
class may be assigned an instance ID, a metric such as intersection over union (loU) may be calculated
with respect to each of the other candidates in the class, and candidates having an loU above some
threshold may be filtered out to remove duplicates.").
Both Ladd ha and Smolyanskiy are in the same field of endeavor of method and systems for
generating ground truth data. It would have been obvious for one ordinary skilled in the a rt before the
effective filing date of present invention to modify Laddha systems and methods for object detection
with Smolyanskiy labeling a cell based on a probability of the cell in the full-range map and a probability
of the cell in the limited-range map. No new functionality would arise from the combination and the
combination would improve usability of Ladd ha by adding the labeling a cell based on a probability of
the cell in the full-range map and a probability of the cell in the limited-range map, one of ordinary skill
in the art would have recognized thatthe results of the combination were predictable.
With respect to dependent claim 19, please see the rejection above with respect to claim 15
which commensurate in scope to claim 19, with claim 15 being drown to a non-transitory computer readable medium, and claim 19 being drawn to a corresponding system.
With respect to dependent claim 20, please see the rejection above with respect to claim 9
which is commensurate in scope to claim 20, with claim 9 being drown to a computer-implemented
method, and claim 20 being drawn to a corresponding system.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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 JEAN PAUL CASS whose telephone number is (571)270-1934. The examiner can normally be reached Monday to Friday 7 am to 7 pm; Saturday 10 am to 12 noon.
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, Scott A. Browne can be reached at 571-270-0151. 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.
/JEAN PAUL CASS/Primary Examiner, Art Unit 3666