US 11,778,167 B1Granted

Method and system for preprocessing optimization of streaming video data

AssigneeInsight Direct USA, Inc.

InventorAmol Ajgaonkar

Life of this patent

Estimating the term…

Loading patent…

Abstract

A method of improving a main output of a main AI model processing first video data includes analyzing incoming video data via a first processing pipeline and analyzing incoming video data via a second processing pipeline. The second processing pipeline includes receiving the incoming video data from the camera; identifying, by a parameter optimization module, first test preprocessing parameters; preprocessing the incoming video data according to the first test preprocessing parameters defined within a test runtime configuration file, wherein the first test preprocessing includes formatting the incoming video data to create first test video data; accessing the first test video data by a test AI model; processing the first test video data by the test AI model to determine a first test output that is indicative of a first test inference dependent upon the first test video data; and comparing the first test output and the main output to a baseline criterion. In response to the first test output satisfying the baseline criterion, the parameter optimization module can alter the first preprocessing parameters in the first runtime configuration file to be similar to the first test preprocessing parameters in the test runtime configuration file and, in response to the first test output failing to satisfy the baseline criterion, replace the first test preprocessing parameters with second test preprocessing parameters that are different from the first test preprocessing parameters and repeat the steps of analyzing the incoming video data via the second preprocessing pipeline.

Claims20

Dependent claims

A method of improving a main output of a main AI model processing first video data, the method comprising:

analyzing incoming video data via a first processing pipeline that comprises:
receiving incoming video data from a camera;
preprocessing the incoming video data according to first preprocessing parameters defined within a first runtime configuration file, wherein the preprocessing includes formatting the incoming video data to create the first video data;
accessing the first video data by the main AI model;
processing the first video data by the main AI model to determine the main output that is indicative of a first inference dependent upon the first video data;
analyzing incoming video data via a second processing pipeline that comprises:
receiving the incoming video data from the camera;
identifying, by a parameter optimization module, first test preprocessing parameters;
preprocessing the incoming video data according to the first test preprocessing parameters defined within a test runtime configuration file, wherein the first test preprocessing includes formatting the incoming video data to create first test video data;
accessing the first test video data by a test AI model;
processing the first test video data by the test AI model to determine a first test output that is indicative of a first test inference dependent upon the first test video data;
comparing the first test output to a baseline criterion;
in response to the first test output satisfying the baseline criterion, altering the first preprocessing parameters in the first runtime configuration file to be similar to the first test preprocessing parameters in the test runtime configuration file;
and
in response to the first test output failing to satisfy the baseline criterion, replacing the first test preprocessing parameters with second test preprocessing parameters that are different from the first test preprocessing parameters and repeating the steps of analyzing the incoming video data via the second preprocessing pipeline.

12 more elements

A system for improving a main output of a main AI model analyzing video data, the system comprising:

a camera configured to collect incoming video data;
a first processing pipeline for analyzing incoming video data, the first processing pipeline comprising:
a first runtime configuration file that includes first preprocessing parameters for formatting the incoming video data;
a first gateway that is configured to receive the incoming video data from the camera and preprocess the incoming video data according to the first preprocessing parameters defined within the first runtime configuration file, wherein the preprocessing of the incoming video data by the first gateway creates first video data;
a first endpoint to which the first video data is published to allow access by the main AI model;
and
the main AI model configured to access the first video data and process the first video data to determine a main output that is indicative of a first inference dependent upon the first video data;
and
a second processing pipeline for analyzing the incoming video data, the second processing pipeline comprising:
a first test runtime configuration file that includes first test preprocessing parameters for formatting the incoming video data;
a parameter optimization module that is configured to receive the incoming video data from the camera and preprocess the incoming video data according to the first test preprocessing parameters defined within the first test runtime configuration file, wherein the test preprocessing includes formatting the incoming video data to create first test video data;
a test endpoint to which the first test video data is published to allow access by a test AI model;
and
the test AI model configured to access the first test video data and process the first test video data to determine a test output that is indicative of a test inference,
wherein the parameter optimization module is configured to access the main output, the test output, and the first runtime configuration file,
wherein the parameter optimization module is configured to compare the test output to a baseline criterion and, in response to the test output satisfying the baseline criterion, alter the first preprocessing parameters in the first runtime configuration file to be similar to the first test preprocessing parameters in the test runtime configuration file,
wherein the parameter optimization module is configured to, in response to the test output failing to satisfy the baseline criterion, replace the first test preprocessing parameters with second test preprocessing parameters that are different from the first test preprocessing parameters and repeating the preprocessing of the incoming video data to create second test video data and the processing of the second test video data by the test AI model.

14 more elements

Dependencies are inferred from the claim text.