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 .
Information Disclosure Statement
The information disclosure statement(s) (IDS) submitted on 05/24/2024 is/are being considered by the Examiner.
Claim Objections
Claim 16 is objected to because of the following informalities: the claim does not appear to have a definitive preamble. Appropriate correction (e.g. placing a semicolon after “comprising”) is required. Furthermore, the claim does not appear to make grammatical sense as it states by saying “one or more labels” then recites “the one or more labels of a labeled synthetic data”. A way to fix this would be to say “using one or more labels, wherein the one or more labels are of a set of labeled synthetic data which are aligned…”
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.
Claim(s) 1, 9, and 10 is/are rejected under 35 U.S.C. 103 as being unpatentable over Romaszko et al, “Learning Direct Optimization for Scene Understanding” (published at https://arxiv.org/abs/1812.07524v1, December 2018) in view of Zhang et al, CN 110070059 A (see attached machine translation).
Regarding claim 1, Romaszko teaches a computer-implemented method, comprising:
encoding, using a scene graph prediction network (see Romaszko page 1, second column, “The latent variables (LVs) z are the scene graph, i.e. the
shape, appearance, position and poses of all objects in the
scene, plus global variables such as the camera and lighting…The Learning Direct Optimization method trains the Prediction Network on data where the current state z does not match the ground truth zGT . This was obtained in two ways: (i) from the initialization network, where based
on x, z0 and g(z0) one can predict zGT , and (ii) by perturbing zGT to produce z0, and learning to predict zGT given x, z0 and g(z0). Note that the training data requirements for Prediction Network”), a set of real data and a set of labeled synthetic data (see page 4, first column, “As our aim is to apply the LiDO method trained on realistic synthetic images for understanding of real images, we apply the same methods to a dataset consisting of real
images with over 750 objects total. The annotated images were manually taken to feature a number of objects of the considered classes at a variety of lighting, viewpoint and object configuration conditions”) into a shared latent space (see page 1, second column, “The latent variables (LVs) z are the scene graph”);
providing labels associated with the real data for updating the scene graph prediction network (see page 5, first column, “For a training image x
with ground truth zGT we obtain a set of image patches P0X extracted at the object detections and the corresponding rendered patches P0R
from the render g(z0). From each pair of patches P0X and P0R p = 1…P and the corresponding z0 the Prediction Network CNN is trained to predict
the object-specific GT variables”).
Romaszko does not expressively teach
the set of real data is a set of unlabeled real data;
aligning one or more labels of the set of labeled synthetic data with one or more instances of real data from the set of unlabeled real data; and
providing the one or more aligned labels associated with the real data for updating the scene graph prediction network.
However, Zhang in a similar invention in the same field of endeavor teaches a method of training a prediction network using a set of real data (see Zhang paragraph [0061]) and a set of labeled synthetic data (see paragraphs [0032] and [0058]) as taught in Romaszko wherein
the set of real data is a set of unlabeled real data (see paragraph [0032]);
the method comprises:
aligning one or more labels of the set of labeled synthetic data with one or more instances of real data from the set of unlabeled real data (see paragraphs [0059]-[0060]); and
providing the one or more aligned labels associated with the real data for updating the prediction network (see paragraph [0061]).
One of ordinary skill in the art before the effective filing date of the invention would have found it obvious to combine the teaching of using unlabeled real data with aligned labels via synthetic data as taught in Zhang with the method taught in Romaszko, the motivation being to eliminate the need for labels in the real data thereby saving time and expense.
Regarding claim 9, Romaszko in view of Zhang teaches all the limitations of claim 1, and further teaches wherein aligning the one or more labels reduces an appearance gap between the labeled synthetic data and the unlabeled real data (see Zhang paragraph [0062]).
Regarding claim 10, Romaszko in view of Zhang teaches all the limitations of claim 1, and further teaches:
receiving an image; and generating a scene graph for the image using the scene graph prediction network (see Romaszko Figure 7 which shows the trained network acting on real images and page 1, second column, “The latent variables (LVs) z are the scene graph, i.e. the shape, appearance, position and poses of all objects in the scene, plus global variables such as the camera and lighting”).
Claim(s) 16, 19, and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Zhang et al, CN 110070059 A (see attached machine translation) in view of Romaszko et al, “Learning Direct Optimization for Scene Understanding” (published at https://arxiv.org/abs/1812.07524v1, December 2018).
Regarding claim 16, Zhang teaches a system comprising one or more processors (see Zhang Figure 1 and paragraph [0042], wherein a processor is implied in order for the system to function) to update a prediction network using one or more labels (see paragraph [0061]), wherein the one or more labels of a set of labeled synthetic data are aligned with a set of unlabeled real data (see paragraphs [0059]-[0060]).
Zhang does not expressively teach wherein the unlabeled real data the labeled synthetic data are encoded into a shared latent space.
However, Romaszko in a similar invention in the same field of endeavor teaches a system comprising one or more processors (see Romaszko Abstract, “Our goal is to explain a single image x with a 3D computer graphics model”) to update a prediction network using one or more labels (see page 5, first column, “For a training image x with ground truth zGT we obtain a set of image patches P0X extracted at the object detections and the corresponding rendered patches P0R from the render g(z0). From each pair of patches P0X and P0R p = 1…P and the corresponding z0 the Prediction Network CNN is trained to predict the object-specific GT variables”) with labeled synthetic data and real data (see page 4, first column, “As our aim is to apply the LiDO method trained on realistic synthetic images for understanding of real images, we apply the same methods to a dataset consisting of real images with over 750 objects total. The annotated images were manually taken to feature a number of objects of the considered classes at a variety of lighting, viewpoint and object configuration conditions”) as taught in Zhang wherein
the unlabeled real data the labeled synthetic data are encoded into a shared latent space (see page 1, second column, “The latent variables (LVs) z are the scene graph”).
One of ordinary skill in the art before the effective filing date of the invention would have found it obvious to combine the teaching of training a prediction network which encodes data into a shared latent space as taught in Romaszko with the system taught in Zhang, the motivation being to utilize the training data for multiple different systems.
Regarding claim 19, Zhang in view of Romaszko teaches all the limitations of claim 16, and further teaches:
receiving an unlabeled image; and generating a scene graph for the image using the scene graph prediction network (see Romaszko Figure 7 which shows the trained network acting on real images and page 1, second column, “The latent variables (LVs) z are the scene graph, i.e. the shape, appearance, position and poses of all objects in the scene, plus global variables such as the camera and lighting”).
Regarding claim 20, Zhang in view of Romaszko teaches all the limitations of claim 16, and further teaches wherein the system comprises at least one of:
a system for performing graphical rendering operations (see Romaszko Abstract); a system for performing simulation operations; a system for performing simulation operations to test or validate autonomous machine applications; a system for performing deep learning operations (see Zhang paragraph [0033]); a system implemented using an edge device; a system incorporating one or more Virtual Machines (VMs); a system implemented at least partially in a data center; or a system implemented at least partially using cloud computing resources.
Allowable Subject Matter
Claims 11-15 are allowed.
The following is an examiner’s statement of reasons for allowance:
Regarding independent claim 11, the prior art made of record fails to teach a processor, comprising:
one or more circuits to:
receive a set of unlabeled real data and a set of labeled synthetic data;
align one or more labels of the synthetic data with the real data in a shared latent space, the shared latent space comprising an encoding of the unlabeled real data and the labeled synthetic data; and
provide the one or more aligned labels of the real data for further training a prediction network.
Zhang, cited above, teaches aligning the labels via training a semantic segmentation network on labeled synthetic data and applying it to unlabeled real data which appears to be a wholly different method than as claimed.
Any comments considered necessary by applicant must be submitted no later than the payment of the issue fee and, to avoid processing delays, should preferably accompany the issue fee. Such submissions should be clearly labeled “Comments on Statement of Reasons for Allowance.”
Claims 2-8, 17, and 18 are 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.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to CASEY L KRETZER whose telephone number is (571)272-5639. The examiner can normally be reached M-F 10:00-7:00 PM Pacific Time.
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, David Payne can be reached at (571)272-3024. 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.
/CASEY L KRETZER/ Primary Examiner, Art Unit 2635
Prosecution Insights