US 12,112,273 B2
Measuring risk within a media scene
Ahsan A. Asghar, Punjab (PK); Freddy Lorge, Vedrin (BE); and David Brian Callies, Spring, TX (US)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Jul. 8, 2020, as Appl. No. 16/923,313.
Prior Publication US 2021/0350256 A1, Nov. 11, 2021
Int. Cl. G06N 5/04 (2023.01); G06N 20/00 (2019.01); G06V 20/40 (2022.01); G06V 20/52 (2022.01)
CPC G06N 5/04 (2013.01) [G06N 20/00 (2019.01); G06V 20/41 (2022.01); G06V 20/52 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method for measuring risk intensity level of a scene captured within media, the computer-implemented method comprising:
generating, by a computer, using a computational scoring model, a score for each risk type, cause, and evidence taxon element described in a risk taxonomy corresponding to a scene class by matching one or more taxon elements with each attribute of each detected object of a set of detected objects within the scene captured by the media, wherein the scene captured within media relates to an area in a building;
tagging, by the computer, the media of the scene with risk type, cause, and evidence scores of each taxon element of the risk taxonomy corresponding to the scene class;
calculating, by the computer, a risk intensity level of the scene based on the risk type, cause, and evidence scores tagged to the media;
detecting, by the computer, a number of targets based on the risk type, cause, and evidence; and
sending, by the computer, a number of alerts based on the risk intensity level of the scene to the number of targets, wherein magnitudes for the number of alerts are determined based on a threshold for the risk intensity level.