Prosecution Insights
Last updated: April 19, 2026
Application No. 18/790,911

REAL-TIME COMMUNICATION CENSORSHIP USING DEEP NEURAL NETWORKS

Non-Final OA §103
Filed
Jul 31, 2024
Examiner
SINHA, SNIGDHA
Art Unit
2619
Tech Center
2600 — Communications
Assignee
Nvidia Corporation
OA Round
1 (Non-Final)
50%
Grant Probability
Moderate
1-2
OA Rounds
2y 6m
To Grant
96%
With Interview

Examiner Intelligence

Grants 50% of resolved cases
50%
Career Allow Rate
3 granted / 6 resolved
-12.0% vs TC avg
Strong +46% interview lift
Without
With
+45.8%
Interview Lift
resolved cases with interview
Typical timeline
2y 6m
Avg Prosecution
26 currently pending
Career history
32
Total Applications
across all art units

Statute-Specific Performance

§101
2.0%
-38.0% vs TC avg
§103
65.6%
+25.6% vs TC avg
§102
16.2%
-23.8% vs TC avg
§112
11.7%
-28.3% vs TC avg
Black line = Tech Center average estimate • Based on career data from 6 resolved cases

Office Action

§103
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 . 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. Claims 1-7, 9-20 are rejected under 35 U.S.C. 103 as being unpatentable over Boudville (US 20240424411) in view of Choi (US 20230259545) and further in view of Lin (CN 113608984). Regarding claim 1, Boudville teaches a computer-implemented method, comprising: Identifying, using the textural content and the at least one neural network, content to be censored (Paragraph 59, GPT detects rude words or gestures); Generating a digital mask corresponding to one or more regions within the chat window containing the content to be censored (Paragraph 49, For printed speech (eg. in a chat box typed by the user), the box's contents can be redacted); Applying one or more visual modifications to the rendered frame based on the digital mask (Paragraph 49, For printed speech (eg. in a chat box typed by the user), the box's contents can be redacted); and Causing a display of the frame with the one or more visual modifications (Paragraph 49, For printed speech (eg. in a chat box typed by the user), the box's contents can be redacted); While Boudville fails to disclose the following, Choi teaches: Extracting textual content from the chat window (Paragraph 8, extract a plurality of texts displayed on a chat window of a message application and collect the texts); Choi and Boudville are both considered to be analogous to the claimed invention because they are in the same field of chat message censoring. Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to have modified Boudville by using Choi and using a neural network to extract text from a chat window. Doing so would allow for identifying the text that may need to be censored. While the combination of Boudville and Choi fails to disclose the following, Lin teaches: Detecting, using at least one neural network, a chat window in a video frame (Page 5, Paragraph 5, convolutional neural network module, is, the analysis model is based on a large number of sample screenshot; and the sample analysis result corresponding to the sample analysis result for deep training learning, can accurately distinguish the target window and non-target window, nontarget window such as office software window office window, and the target window such as game application window, shopping website, chat software window); Lin and the combination of Boudville and Choi are both considered to be analogous to the claimed invention because they are in the same field of chat message censoring. Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to have modified the combination of Boudville and Choi by using Lin and using a neural network to detect a chat box in a video frame. Doing so would allow for efficiently identifying the location of any content that needs to be censored. System claim 9 corresponds to apparatus claim 1, where the communication interface of claim 9 is the chat window of claim 1. Therefore, claim 9 is rejected for the same rationale as above. Regarding claim 2, the combination of Boudville, Choi, and Lin teaches the computer-implemented method of claim 1, wherein the one or more visual modifications are based at least on a modification type associated with at least one modification criterion (Boudville, Paragraph 49, The choice of which action to be taken against written speech can vary, perhaps depending on the age of the recipients. Eg. for kids, a stricter action can be done. Or depending on the severity of the (eg) swearing. Or depending on the audience). System claim 10 corresponds to apparatus claim 2. Therefore, claim 9 is rejected for the same rationale as above. System claim 17 corresponds to apparatus claim 2. Independent claims 1 and 16 recite different limitations but are taught by the same combination of references. Therefore, claim 17 is rejected for the same rationale as above. Regarding claim 3, the combination of Boudville, Choi, and Lin teaches the computer-implemented method of claim 1. While the combination as presented previously fails to disclose the following, Choi further teaches: Wherein the chat window allows interactive communication between two or more users (Paragraph 4, information on texts displayed on a chat window screen of the message application; Figure 5A). Choi and the combination of Boudville and Lin are both considered to be analogous to the claimed invention because they are in the same field of chat message censoring. Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to have modified the combination of Boudville and Lin by using Choi and use a chat window between two or more users. Doing so would allow for moderating live chat and ensuring both parties view appropriate content. System claim 11 corresponds to apparatus claim 3, where the communication interface of claim 11 is the chat window of claim 3. Therefore, claim 11 is rejected for the same rationale as above. System claim 18 corresponds to apparatus claim 3, where the communication interface of claim 18 is the chat window of claim 3. Independent claims 1 and 16 recite different limitations but are taught by the same combination of references. Therefore, claim 18 is rejected for the same rationale as above. Regarding claim 4, the combination of Boudville, Choi, and Lin teaches the computer-implemented method of claim 1, wherein the applying the one or more visual modifications comprises at least one or blurring the content to be censored, redacting the content to be censored, or replacing the content to be censored with pre-approved content (Boudville, Paragraph 66, FIG. 6 shows examples. Item 61 is the Predator's original text—“You are such a dirty girl”. Item 62 is a modded strikethrough—“You are such a custom-character”. We are using the GPT to detect “dirty girl”. Item 63 has a fully cancelled out text—“You are such a xxxxxxxxx”. While item 64 has the Predator's text altered to poke fun at the Predator himself—“I am so DUMB!!!”). System claim 12 corresponds to apparatus claim 4. Therefore, claim 12 is rejected for the same rationale as above. System claim 19 corresponds to apparatus claim 4. Independent claims 1 and 16 recite different limitations but are taught by the same combination of references. Therefore, claim 19 is rejected for the same rationale as above. Regarding claim 5, the combination of Boudville, Choi, and Lin teaches the computer-implemented method of claim 1, further comprising, executing a post-processing engine to apply the one or more visual modifications to the frame (Boudville, Paragraph 49, These changes might require a modification of the GUI widgets in which the written text is shown). Regarding claim 6, the combination of Boudville, Choi, and Lin teaches the computer-implemented method of claim 1, further comprising, analyzing, using a large language model (LLM), the textual content to determine if the textual content is of a specific sentiment (Boudville, Paragraph 38, We use GPT as synonymous with accessing a Large Language Model (LLM); Paragraph 64, the GPT detects rude written words). System claim 13 corresponds to apparatus claim 6. Therefore, claim 13 is rejected for the same rationale as above. Regarding claim 7, the combination of Boudville, Choi, and Lin teaches the computer-implemented method of claim 6, further comprising, in response to determining that the textural content is of the specific sentiment, categorizing the specific sentiment into a plurality of levels of severity, and wherein the applying the one or more visual modifications is based on the levels of severity (Boudville, Paragraph 49, The choice of which action to be taken against written speech can vary, perhaps depending on the age of the recipients. Eg. for kids, a stricter action can be done. Or depending on the severity of the (eg) swearing). System claim 14 corresponds to apparatus claim 7. Therefore, claim 14 is rejected for the same rationale as above. Regarding claim 15, the combination of Boudville, Choi, and Lin teaches the processor of claim 9, wherein the processor is included in a system comprising at least one of: A system for performing simulation operations; A system for performing simulation operations to test or validate autonomous machine applications; A system for performing digital twin operations; A system for performing light transport simulation; A system for rendering graphical output (Boudville, Paragraph 55, These changes might require a modification of the GUI widgets in which the written text is shown); A system for performing deep learning operations; A system implemented using an edge device; A system for generating or presenting virtual reality (VR) content (Boudville, Paragraph 14, Our work can be applied to the Metaverse. By this, we take the Metaverse to be Virtual Reality (VR) plus the use of avatars within VR); A system for generating or presenting augmented reality (AR) content; A system for generating or presenting mixed reality (MR) content; A system incorporating one or more virtual machines (VMs); A system implemented at least partially in a data center; A system for performing hardware testing using simulation; A system for synthetic data generation; A system for performing generative AI operations; A system implemented using one or more large language model (LLMs) (Boudville, Paragraph 38, We use GPT as synonymous with accessing a Large Language Model (LLM)); A system implemented using one or more vision language model (VLMs); A collaborative content creation platform for 3D assets; or A system implemented at least partially using cloud computing resources. System claim 20 corresponds to system claim 15. Independent claims 9 and 16 recite different limitations but are taught by the same combination of references. Therefore, claim 20 is rejected for the same rationale as above. Regarding claim 16, Boudville teaches a system, comprising: One or more processing units to apply one or more visual modifications and to cause a display of the video frame with the one or more visual modification, wherein the one or more visual modifications are generated for one or more regions within the communication interface that are determined to contain content that is restricted from being displayed (Paragraph 59, GPT detects rude words or gestures; Paragraph 49, For printed speech (eg. in a chat box typed by the user), the box's contents can be redacted). While Boudville fails to disclose the following, Choi teaches: The content determined in part by analyzing the textual content in the communication interface (Paragraph 8, extract a plurality of texts displayed on a chat window of a message application and collect the texts). Choi and Boudville are both considered to be analogous to the claimed invention because they are in the same field of chat message censoring. Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to have modified Boudville by using Choi to extract text from a chat window. Doing so would allow for identifying the text that may need to be censored. While the combination of Boudville and Choi fails to disclose the following, Lin teaches: A communication interface detected in a video frame (Page 5, Paragraph 5, convolutional neural network module, is, the analysis model is based on a large number of sample screenshot; and the sample analysis result corresponding to the sample analysis result for deep training learning, can accurately distinguish the target window and non-target window, nontarget window such as office software window office window, and the target window such as game application window, shopping website, chat software window); Lin and the combination of Boudville and Choi are both considered to be analogous to the claimed invention because they are in the same field of chat message censoring. Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to have modified the combination of Boudville and Choi by using Lin and detect a chat box in a video frame. Doing so would allow for efficiently identifying the location of any content that needs to be censored. Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Boudville in view of Choi and further in view of Lin as applied to claims 1-7, 9-20 above and further in view of Pedersen (US 20200142999). Regarding claim 8, the combination of Boudville, Choi, and Lin teaches the computer-implemented method of claim 1. While the combination fails to disclose the following, Pedersen teaches: Wherein the identifying comprises performing reverse substitution from non-alphabetic characters in the textural content to identify offensive content (Paragraph 3, such as by intentionally misspelling words (e.g., misspelled profanity, hate speech, etc.) or using symbols in lieu of letters of the alphabet to disguise toxic language while still conveying its meaning to viewing users). Pedersen and the combination of Boudville, Choi, and Lin are both considered to be analogous to the claimed invention because they are in the same field of chat message censoring. Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to have modified the combination of Boudville, Choi, and Lin by using Pedersen and identifying non-alphabetic character substitution to identify offensive content. Doing so would allow detecting creative ways of circumventing automated detection systems (Pedersen, Paragraph 3). Conclusion Any inquiry concerning this communication or earlier communications from the examiner should be directed to SNIGDHA SINHA whose telephone number is (571)272-6618. The examiner can normally be reached Mon-Fri. 12pm-8pm. 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 at 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 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. /SNIGDHA SINHA/Examiner, Art Unit 2619 /JASON CHAN/Supervisory Patent Examiner, Art Unit 2619
Read full office action

Prosecution Timeline

Jul 31, 2024
Application Filed
Mar 03, 2026
Non-Final Rejection — §103 (current)

Precedent Cases

Applications granted by this same examiner with similar technology

Patent 12567216
AUGMENTED-REALITY-INTERFACE CONFLATION IDENTIFICATION
2y 5m to grant Granted Mar 03, 2026
Patent 12406339
MACHINE LEARNING DATA AUGMENTATION USING DIFFUSION-BASED GENERATIVE MODELS
2y 5m to grant Granted Sep 02, 2025
Study what changed to get past this examiner. Based on 2 most recent grants.

AI Strategy Recommendation

Get an AI-powered prosecution strategy using examiner precedents, rejection analysis, and claim mapping.
Powered by AI — typically takes 5-10 seconds

Prosecution Projections

1-2
Expected OA Rounds
50%
Grant Probability
96%
With Interview (+45.8%)
2y 6m
Median Time to Grant
Low
PTA Risk
Based on 6 resolved cases by this examiner. Grant probability derived from career allow 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