Point cloud annotation device, method, and program
Annotation can be easily performed on a three-dimensional point cloud and a working time can be reduced. An interface unit 22 displays a point cloud indicating a three-dimensional point on an object, and receives designation of a three-dimensional point indicating an annotation target object and designation of a three-dimensional point not indicating the annotation target object. A candidate cluster calculation unit 32 calculates a value of a predetermined evaluation function indicating a likelihood of a point cloud cluster being the annotation target object based on the designation of a three-dimensional point for point cloud clusters obtained by clustering the point clouds. A cluster selection and storage designation unit 34 causes the interface unit 22 to display the point cloud clusters in descending order of the value of the evaluation function, and receives a selection of a point cloud cluster to be annotated. An annotation execution unit 36 executes annotation indicating the annotation target object for each three-dimensional point included in the selected point cloud cluster.
1. A point cloud annotation device comprising a processor configured to execute operations comprising:
displaying point clouds, wherein the point clouds comprise a plurality of three-dimensional points;
interactively receiving designation of a first three-dimensional point of the plurality of three-dimensional points and a second three-dimensional point of the plurality of three-dimensional points, wherein the first three-dimensional point indicates an annotation target object and represents a positive example, and the second three-dimensional point excludes indication of the annotation target object and represents a negative example;
generating a plurality of point cloud clusters in the point clouds;
determining a value of a predetermined evaluation function of a point cloud cluster of the plurality of point cloud clusters, wherein the value of the predetermined evaluation function indicates a likelihood of a point cloud cluster being the annotation target object, based on the first three-dimensional point as the positive example and the second three-dimensional point as the negative example;
causing an interface processor to display the plurality of point cloud clusters in descending order of the value of the predetermined evaluation function;
interactively receiving a selection of the point cloud cluster of the plurality of point cloud clusters to be annotated; and
executing an annotation indicating the annotation target object for each three-dimensional point included in the point cloud cluster.
2. The point cloud annotation device according to claim 1 , wherein the predetermined evaluation function outputs a greater value as the point cloud cluster includes more three-dimensional points indicating the annotation target object, and the predetermined evaluation function further outputs a smaller value as the point cloud cluster includes more three-dimensional points not indicating the annotation target object.
3. The point cloud annotation device according to claim 2 , wherein the point cloud cluster is obtained from a result of clustering the point cloud using a plurality of clustering schemes.
4. The point cloud annotation device according to claim 2 , the processor further configured to execute operations comprising:
repeating a set of operations until a predetermined condition is satisfied, wherein the set of operations including:
the receiving designation,
the determining a value of the predetermined evaluation function,
the receiving a selection of a point cloud cluster to be annotated, and
the executing one or more annotations based on the received selection.
5. The point cloud annotation device according to claim 1 , wherein the point cloud cluster is obtained from a result of clustering the point clouds using a plurality of clustering schemes.
6. The point cloud annotation device according to claim 5 , the processor further configured to execute operations comprising:
repeating a set of operations until a predetermined condition is satisfied, wherein the set of operations including:
the receiving designation,
the determining a value of the predetermined evaluation function,
the receiving a selection of a point cloud cluster to be annotated, and
the executing one or more annotations based on the received selection.
7. The point cloud annotation device according to claim 1 , the processor further configured to execute operations comprising:
repeating a set of operations until a predetermined condition is satisfied, wherein the set of operations including:
the receiving designation,
the determining a value of the predetermined evaluation function,
the receiving a selection of a point cloud cluster to be annotated, and
the executing one or more annotations based on the received selection.
8. A point cloud annotation method comprising:
displaying point clouds, wherein the point clouds comprise a plurality of three-dimensional points;
interactively receiving designation of a first three-dimensional point of the plurality of three-dimensional points and a second three-dimensional point of the plurality of three-dimensional points, wherein the first three-dimensional point indicates an annotation target object and represents a positive example, and the second three-dimensional point excludes indication of the annotation target object and represents a negative example;
generating a plurality of point cloud clusters in the point clouds;
determining a value of a predetermined evaluation function of a point cloud cluster of the plurality of point cloud clusters, indicating a likelihood of a point cloud cluster being the annotation target object, based on the first three-dimensional point as the positive example and the second three-dimensional point as the negative example;
causing the interactive processor to display the plurality of point cloud clusters in descending order of the value of the predetermined evaluation function;
interactively receiving a selection of the point cloud cluster of the plurality of point cloud clusters to be annotated; and
executing an annotation indicating the annotation target object for each three-dimensional point included in the point cloud cluster.
9. The point cloud annotation method according to claim 8 , wherein the predetermined evaluation function outputs a greater value as the point cloud cluster includes more three-dimensional points indicating an annotation target object, and the predetermined evaluation function further outputs a smaller value as the point cloud cluster includes more three-dimensional points not indicating the annotation target object.
10. The point cloud annotation method according to claim 9 , wherein the point cloud cluster is obtained from a result of clustering the point cloud using a plurality of clustering schemes.
11. The point cloud annotation method according to claim 9 , further comprising:
repeating a set of operations until a predetermined condition is satisfied, wherein the set of operations including:
the receiving designation,
the determining a value of the predetermined evaluation function,
the receiving a selection of a point cloud cluster to be annotated, and
the executing one or more annotations based on the received selection.
12. The point cloud annotation method according to claim 8 , wherein the point cloud cluster is obtained from a result of clustering the point cloud using a plurality of clustering schemes.
13. The point cloud annotation method according to claim 12 , further comprising:
repeating a set of operations until a predetermined condition is satisfied, wherein the set of operations including:
the receiving designation,
the determining a value of the predetermined evaluation function,
the receiving a selection of a point cloud cluster to be annotated, and
the executing one or more annotations based on the received selection.
14. The point cloud annotation method according to claim 8 , further comprising:
repeating a set of operations until a predetermined condition is satisfied, wherein the set of operations including:
the receiving designation,
the determining a value of the predetermined evaluation function,
the receiving a selection of a point cloud cluster to be annotated, and
the executing one or more annotations based on the received selection.
15. A computer-readable non-transitory recording medium storing computer-executable program instructions that when executed by a processor cause a computer system to execute operations comprising:
displaying point clouds, wherein the point clouds comprise a plurality of three-dimensional points;
interactively receiving designation of a first three-dimensional point of the plurality of three-dimensional points and a second three-dimensional point of the plurality of three-dimensional points, wherein the first three-dimensional point indicates an annotation target object and represents a positive example, and the second three-dimensional point excludes indication of the annotation target object and represents a negative example;
generating a plurality of point cloud clusters in the point clouds;
determining a value of a predetermined evaluation function of a point cloud cluster of the plurality of point cloud clusters, wherein the value of the predetermined evaluation function indicates a likelihood of a point cloud cluster being the annotation target object, based on the first three-dimensional point as the positive example and the second three-dimensional point as the negative example;
interactively receiving a selection of the point cloud cluster of the plurality of point cloud clusters to be annotated; and
executing, an annotation indicating the annotation target object for each three-dimensional point included in the point cloud cluster.
16. The computer-readable non-transitory recording medium of claim 15 , wherein the predetermined evaluation function outputs a greater value as the point cloud cluster includes more three-dimensional points indicating an annotation target object, and the predetermined evaluation function further outputs a smaller value as the point cloud cluster includes more three-dimensional points not indicating the annotation target object.
17. The computer-readable non-transitory recording medium of claim 16 , wherein the point cloud cluster is obtained from a result of clustering the point cloud using a plurality of clustering schemes.
18. The computer-readable non-transitory recording medium of claim 15 , wherein the point cloud cluster is obtained from a result of clustering the point cloud using a plurality of clustering schemes.
19. The computer-readable non-transitory recording medium of claim 18 , the computer-executable program instructions that when executed by the processor further cause the computer system to execute operations comprising:
repeating a set of operations until a predetermined condition is satisfied, wherein the set of operations including:
the receiving designation,
the determining a value of the predetermined evaluation function,
the receiving a selection of a point cloud cluster to be annotated, and
the executing one or more annotations based on the received selection.
20. The computer-readable non-transitory recording medium of claim 15 , the computer-executable program instructions that when executed by the processor further cause the computer system to execute operations comprising:
repeating a set of operations until a predetermined condition is satisfied, wherein the set of operations including:
the receiving designation,
the determining a value of the predetermined evaluation function,
the receiving a selection of a point cloud cluster to be annotated, and
the executing one or more annotations based on the received selection.