US 12,293,586 B2
Computer-readable recording medium storing person identification determination program, person identification determination method, and information processing apparatus
Shun Kohata, Setagaya (JP)
Assigned to FUJITSU LIMITED, Kawasaki (JP)
Filed by FUJITSU LIMITED, Kawasaki (JP)
Filed on Jun. 10, 2022, as Appl. No. 17/837,504.
Claims priority of application No. 2021-168431 (JP), filed on Oct. 13, 2021.
Prior Publication US 2023/0113045 A1, Apr. 13, 2023
Int. Cl. G06V 20/52 (2022.01); G06T 7/73 (2017.01); G06V 10/82 (2022.01); G06V 40/16 (2022.01)
CPC G06V 20/52 (2022.01) [G06T 7/73 (2017.01); G06V 40/172 (2022.01); G06T 2207/20081 (2013.01); G06T 2207/30196 (2013.01)] 5 Claims
OG exemplary drawing
 
1. A non-transitory computer-readable recording medium storing a determination program for causing a computer to execute processing comprising:
acquiring a plurality of pieces of image data captured by each of a plurality of cameras;
specifying a position of a person included in each of the plurality of pieces of image data by a first index different for each of the plurality of cameras;
specifying the position of the person specified by the first index by a second index common to the plurality of cameras;
determining whether the person included in each of the plurality of pieces of image data is the same person on the basis of the specified position of the person that uses the second index; and
generating, as learning data used for machine learning of a machine learning model for person identification, which executes two-class classification that identifies whether or not persons who appear in a plurality of pieces of input data are the same person and constitutes a tracking system for the same person together with a person detection model, the learning data obtained by assigning a correct answer label that indicates that the same person appears to the plurality of pieces of image data determined to be the same person,
wherein the learning data is obtained by:
calculating a conversion coefficient from a first coordinate system for the plurality of cameras, which is used for the first index, to a second coordinate system for the person, which is used for the second index;
detecting first area information in which the position of the person included in each piece of image data captured at the same time by the plurality of cameras is specified in the first coordinate system;
converting the first area information into second area information of the second coordinate system by using the conversion coefficient; and
acquiring a data set of pieces of the second area information positioned at equivalent coordinates in the second coordinate system among the plurality of cameras.