US 11,893,683 B2
Systems and methods for artificial intelligence (AI) three-dimensional modeling
Hoa Ton-That, Glastonbury, CT (US); Ryan M. Scanlon, East Hampton, CT (US); and Douglas L. Roy, Plantsville, CT (US)
Assigned to The Travelers Indemnity Company, Hartford, CT (US)
Filed by The Travelers Indemnity Company, Hartford, CT (US)
Filed on Feb. 23, 2023, as Appl. No. 18/173,260.
Application 18/173,260 is a continuation of application No. 17/353,008, filed on Jun. 21, 2021, granted, now 11,625,893.
Prior Publication US 2023/0206552 A1, Jun. 29, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06T 17/00 (2006.01); G06N 20/00 (2019.01); G06N 5/04 (2023.01); G06T 7/13 (2017.01); G06T 7/73 (2017.01); G06T 3/00 (2006.01); G06T 7/60 (2017.01); G06T 7/12 (2017.01); G06F 18/2413 (2023.01); H04N 23/698 (2023.01)
CPC G06T 17/00 (2013.01) [G06F 18/2413 (2023.01); G06N 5/04 (2013.01); G06N 20/00 (2019.01); G06T 3/00 (2013.01); G06T 7/12 (2017.01); G06T 7/13 (2017.01); G06T 7/60 (2013.01); G06T 7/73 (2017.01); G06T 2207/30244 (2013.01); H04N 23/698 (2023.01)] 20 Claims
OG exemplary drawing
 
1. An Artificial Intelligence (AI) three-dimensional modeling method, comprising:
receiving, by an electronic processing device, at least one image descriptive of a room as viewed from a plurality of perspectives;
identifying, by the electronic processing device and from the at least one image, a plurality of corners of the room that are represented at a plurality of respective corner portions of the at least one image;
identifying, by the electronic processing device and based on the identified plurality of corners of the room, at least one camera location within the room;
identifying, by the electronic processing device and based on the identified plurality of corners of the room and the identified at least one camera location within the room, a plurality of walls of the room that are represented by a plurality of respective wall portions of the at least one image;
segmenting, by the electronic processing device and based on the identified plurality of walls of the room, the image into the plurality of respective wall portions of the at least one image;
normalizing, by the electronic processing device, the plurality of respective wall portions of the at least one image;
defining, by the electronic processing device, an array of coordinates for each of the plurality of respective wall portions of the at least one image;
projecting, by the electronic processing device and utilizing the array of coordinates for each of the plurality of respective wall portions of the at least one image, the plurality of respective wall portions of the at least one image into a three-dimensional model;
identifying, by the electronic processing device, an object within the at least one image;
classifying, by the electronic processing device, the object, thereby defining a classification of the object;
identifying, by the electronic processing device and based on the classifying of the object, a rule for the object, wherein the identifying comprises querying, utilizing data descriptive of the classification of the object, a plurality of stored rules and identifying, from the plurality of stored rules, the rule for the object;
calculating, by the electronic processing device and based on the rule, an assessment for the room; and
outputting, by the electronic processing device, an indication of the assessment.