US 11,836,981 B2
Method for assisting real-time monitoring of at least one person on sequences of images
Benoît Pelletier, Saint Etienne de Crossey (FR); Edoh Kodjo, Grenoble (FR); Baptiste Metge, Grenoble (FR); Rémi Druilhe, Saint Martin d'Hares (FR); Cécile Boukamel-Donnou, Echirolles (FR); and Nicolas Arnaise, Echirolles (FR)
Assigned to BULL SAS, Les Clayes-sous-Bois (FR)
Filed by BULL SAS, Les Clayes-sous-bois (FR)
Filed on Jun. 30, 2021, as Appl. No. 17/364,782.
Claims priority of application No. 20305764 (EP), filed on Jul. 6, 2020.
Prior Publication US 2022/0004768 A1, Jan. 6, 2022
Int. Cl. G06V 20/40 (2022.01); G06V 10/56 (2022.01); G06F 18/21 (2023.01)
CPC G06V 20/40 (2022.01) [G06F 18/2178 (2023.01); G06V 10/56 (2022.01)] 12 Claims
OG exemplary drawing
 
1. A method for assisting real-time monitoring of at least one person of interest, implemented by a computer, comprising:
obtaining a reference tracklet comprising at least one sequence of images of said at least one person of interest;
obtaining in real time a new tracklet comprising at least one new sequence of images of a new person;
analyzing said new tracklet to determine whether said new tracklet may be associated with said at least one person of interest, comprising
calculating a distance between a signature of said reference tracklet and a signature of said new tracklet; and
when said distance is less than or equal to a predetermined threshold, filtering said new tracklet with at least one parameter relating to said reference tracklet; and,
when said new tracklet may be associated with said at least one person of interest, validating whether said new tracklet is associated with said at least one person of interest.