US 12,266,164 B2
Classification parallelization architecture
William S. Bowman, Melbourne, FL (US); Sean Wagoner, West Melbourne, FL (US); Andrew D. Falendysz, Grant, FL (US); Matthew D. Summer, Melbourne, FL (US); Kevin Makovy, West Melbourne, FL (US); Jeffrey S. Cooper, Centreville, VA (US); and Brad Truesdell, Indialantic, FL (US)
Assigned to Tomahawk Robotics, Inc., Melbourne, FL (US)
Filed by Tomahawk Robotics, Inc., Melbourne, FL (US)
Filed on Jul. 12, 2024, as Appl. No. 18/772,099.
Application 18/772,099 is a continuation of application No. 18/454,055, filed on Aug. 22, 2023, granted, now 12,067,768.
Application 18/454,055 is a continuation of application No. 17/571,081, filed on Jan. 7, 2022, granted, now 11,776,247, issued on Oct. 3, 2023.
Prior Publication US 2024/0378880 A1, Nov. 14, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G06V 10/764 (2022.01); G05B 13/02 (2006.01); G06V 10/70 (2022.01); G06V 10/94 (2022.01); G06V 10/96 (2022.01); G06V 20/56 (2022.01)
CPC G06V 10/96 (2022.01) [G05B 13/0265 (2013.01); G06V 10/764 (2022.01); G06V 10/87 (2022.01); G06V 10/955 (2022.01); G06V 20/56 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A system for generating composite images that include object identifiers determined based on processing source images through multiple machine learning models, the system comprising:
one or more processors; and
a non-transitory computer-readable storage medium storing instructions, which when executed by the one or more processors cause the one or more processors to perform operations comprising:
receiving a plurality of images from a recording device;
accessing a plurality of machine learning models for processing the plurality of images;
inputting a first image into each of the plurality of machine learning models;
receiving output from a first machine learning model of the plurality of machine learning models;
determining that a subset of images of the plurality of images has been input into a second machine learning model of the plurality of machine learning models;
inputting a last image of the subset of images into the first machine learning model while refraining from inputting other images of the subset of images into the first machine learning model;
receiving from each machine learning model one or more indicators indicating one or more objects identified in corresponding images of the plurality of images;
generating a set of composite images for the plurality of images with the one or more indicators from each machine learning model; and
generating an output video stream comprising the set of composite images to be displayed in chronological order.