Prior art — US Patent 11753046

US patent 11753046, titled "System and method of predicting human interaction with vehicles," does not explicitly list a separate "Cited Patents" section within the provided authoritative text. However, the "CROSS REFERENCE TO RELATED APPLICATIONS" section details the patent family lineage, incorporating the contents of several earlier applications and granted patents by reference. For the purpose of identifying "patent citations" as requested, this analysis will focus on the two granted patents within this family chain, as their disclosures are incorporated by reference and predate the filing date of US11753046.

It is important to note that, typically, parent applications/patents in a valid continuation chain, where the claims of the later application are fully supported by the disclosure of the earlier applications and properly claim priority, are not considered prior art under 35 U.S.C. § 102 against the claims of the later application. Instead, they establish the priority date for the claimed invention. However, if claims in the later application introduce new matter not supported by the earlier applications, or if priority is not properly maintained for certain claims, then the earlier documents (including granted patents) could potentially be considered prior art against such claims. Given the instruction to identify what "potentially anticipates," this analysis proceeds under the theoretical condition that a proper priority chain might not cover all aspects, or to illustrate the overlap in inventive disclosure.

The core inventive features of US11753046, as outlined in its "Summary of the Invention" and "Abstract," include:

Receiving image/video data of a road scene from a participant's perspective (e.g., vehicle sensors).
Generating stimulus data (original or altered image/video).
Transmitting stimulus data to a user interface for human observation.
Receiving response data from human observers (actions, likelihoods of actions of road participants).
Aggregating response data into statistical data.
Creating a model based on statistical data, often using supervised learning algorithms (e.g., neural networks).
Applying the model to new image/video data (e.g., live road scenes).
Generating a prediction of user behavior (e.g., intention, awareness) based on the model.

Below are the patent citations found in the "CROSS REFERENCE TO RELATED APPLICATIONS" section that are themselves granted patents:

1. U.S. Pat. No. 10,614,344

Full Citation: U.S. Pat. No. 10,614,344 B2, titled "System and method of predicting human interaction with vehicles," issued on April 7, 2020.
Publication/Filing Date: Issued: April 7, 2020. This patent is a continuation of U.S. patent application Ser. No. 16/512,560, filed on July 16, 2019. The earliest priority date claimed by this patent through its chain of applications (and shared by US11753046) is July 5, 2017 (U.S. Provisional Application No. 62/528,771).
Brief Description: As a direct parent in the continuation chain, the disclosure of US10614344 is substantially identical to that of US11753046. It describes a system and method for predicting human interaction with vehicles by training a machine learning model using human observer responses to stimulus data derived from road scene images/videos. The system involves generating stimulus data, collecting human input regarding the likely actions or mental states of road users (pedestrians, cyclists, other vehicles), aggregating this data into statistical summaries, and training a supervised learning algorithm (such as random forest, SVM, or various neural networks) to predict behavior in live road scenes.
Potential Anticipation (35 U.S.C. § 102): The disclosure of US10614344 B2 would, by virtue of its highly similar or identical content as a parent application incorporated by reference, anticipate all of the core inventive features described in the "Summary of the Invention" of US11753046. This includes the steps of receiving image/video, generating stimulus data, collecting human responses, aggregating statistical data, training a model with supervised learning, applying the model to live data, and predicting user behavior. Assuming proper priority is claimed for all claims in US11753046, this anticipation would generally be overcome by the priority claim, meaning US10614344 would not be considered prior art against claims entitled to its priority date. However, any potential claims in US11753046 that might contain "new matter" not fully disclosed in US10614344, and thus not entitled to its priority date, could theoretically be anticipated by the disclosure of US10614344.

2. U.S. Pat. No. 10,402,687

Full Citation: U.S. Pat. No. 10,402,687 B2, titled "System and method of predicting human interaction with vehicles," issued on September 3, 2019.
Publication/Filing Date: Issued: September 3, 2019. This patent is a continuation of U.S. patent application Ser. No. 15/830,549, filed on December 4, 2017. The earliest priority date claimed by this patent through its chain of applications (and shared by US11753046) is July 5, 2017 (U.S. Provisional Application No. 62/528,771).
Brief Description: Similar to US10614344, US10402687 is also a direct parent in the continuation chain for US11753046. Its disclosure is therefore substantially identical, covering the fundamental system and method for predicting human interaction with vehicles. This includes the process of gathering road scene imagery, creating modified stimuli for human evaluation, collecting human judgments on road user behavior (intentions, awareness), processing these responses into statistical data, and using this data to train a machine learning model for real-time prediction in autonomous driving or other vehicle systems.
Potential Anticipation (35 U.S.C. § 102): The disclosure of US10402687 B2, being a direct parent in the patent family with incorporated by reference content, would likewise anticipate all of the core inventive features summarized in US11753046. This includes the methodologies for generating and presenting stimuli, gathering and analyzing human responses (e.g., content, response time, eye tracking), and utilizing this aggregate human intelligence to train a predictive model (including various neural network architectures) for real-world application in vehicles. As with US10614344, assuming proper priority is maintained for all claims in US11753046, US10402687 would not typically be considered prior art against claims entitled to its priority date. However, any claims in US11753046 that represent "new matter" not fully supported by US10402687 and thus not eligible for its priority date could be anticipated by the disclosure within US10402687.