Attention target estimating device, and attention target estimating method
An objective of the present disclosure is to enable estimation of an attention target of communication parties even if there are a plurality of objects having the same appearance in a space. An attention target estimation device 10 according to the present disclosure acquires a first-person point-of-view video IM i captured from a perspective of a person and a line-of-sight position g i of a person when the first-person point-of-view video IM i is captured, identifies positions in a 3D space of objects 31, 32 , and 33 extracted from the first-person point-of-view video IM i , and determines an object close to the line-of-sight position g i of the person among the objects 31, 32 , and 33 included in the first-person point-of-view video IM i as the attention target of the person.
1. An attention target estimation device comprising:
a processor; and
a storage medium having computer program instructions stored thereon, when executed by the processor, perform to:
acquire a first-person point-of-view video captured from a perspective of a person and information of a line of sight of the person when the first-person point-of-view video is captured;
identify a position of an object extracted from the first-person point-of-view video in a three-dimensional (3D) space;
determine an object close to the line of sight of the person among objects included in the first-person point-of-view video as an attention target of the person;
estimate an object of the attention target for each of a plurality of persons communicating in the same space; and
determine whether joint attention has occurred based on the objects of the attention target of the plurality of persons.
2. The attention target estimation device according to claim 1 , wherein the computer program instructions further perform to determine an object close to the line of sight of the person and close to the line of sight of a second person different from the person among the objects included in the first-person point-of-view video as the attention target of the person.
3. The attention target estimation device according to claim 1 , wherein the computer program instructions further perform to acquire determination results of attention targets of the plurality of persons, and when objects or positions in the 3D space of the attention targets of the plurality persons are similar, determining that joint attention has occurred.
4. The attention target estimation device according to claim 1 , wherein
the first-person point-of-view video is a moving image, and
an attention target previously determined as an attention target among the objects included in the first-person point-of-view video is determined as the attention target of the person.
5. The attention target estimation device according to claim 1 , wherein
the information of the line of sight is a position in the first-person point-of-view video, and
the line of sight of the person is a straight line starting at a presence position of the person in a real space and ending at the position indicated by the information of the line of sight in the first-person point-of-view video.
6. An attention target estimation method performed by an attention target estimation device, the attention target estimation method comprising:
by a processor, acquiring a first-person point-of-view video captured from a perspective of a person and information of a line of sight of the person when the first-person point-of-view video is captured;
by the processor, identifying a position in a three-dimensional (3D) space of an object extracted from the first-person point-of-view video;
by the processor, determining an object close to the line of sight of the person among objects included in the first-person point-of-view video as an attention target of the person;
by the processor, estimating an object of the attention target for each of a plurality of persons communicating in the same space; and
by the processor, determining whether joint attention has occurred based on the objects of the attention target of the plurality of persons.
7. The attention target estimation method to claim 6 further comprising determining an object close to the line of sight of the person and close to the line of sight of a second person different from the person among the objects included in the first-person point-of-view video as the attention target of the person.
8. The attention target estimation method according to claim 6 further comprising acquiring determination results of attention targets of the plurality of persons, and when objects or positions in the 3D space of the attention targets of the plurality persons are similar, determining that joint attention has occurred.
9. The attention target estimation method according to claim 6 , wherein
the first-person point-of-view video is a moving image, and
an attention target previously determined as an attention target among the objects included in the first-person point-of-view video is determined as the attention target of the person.
10. The attention target estimation method according to claim 6 , wherein
the information of the line of sight is a position in the first-person point-of-view video, and
the line of sight of the person is a straight line starting at a presence position of the person in a real space and ending at the position indicated by the information of the line of sight in the first-person point-of-view video.