Image processing system, image processing method, and program

1. An image processing system, comprising at least one processor configured to:

acquire taken images that have been taken by a camera, which is movable in a real space;

acquire, based on changes in position of a feature point cloud in the taken images, observation space information including three-dimensional coordinates of the feature point cloud in an observation space;

acquire, based on machine learning data on a feature of an object, additional information on a feature of a photographed object shown in the taken images;

integrate the observation space information and the additional information;

wherein the additional information includes two-dimensional feature amount information in which a position of the photographed object in the taken images and a feature amount regarding the photographed object are associated with each other,

estimate a position of the camera based on the changes in position of the feature point cloud, and set an observation viewpoint in the observation space based on a result of the estimation, and

execute processing based on a result of comparison between two-dimensional observation information, which shows a view of the observation space as seen from the observation viewpoint, and the two-dimensional feature amount information;

wherein the feature amount includes a depth of the photographed object estimated based on the machine learning data,

wherein, in the two-dimensional observation information, a position of the feature point cloud in a two-dimensional space, and a depth of the feature point cloud in the observation space are associated with each other, and

wherein the at least one processor is configured to set a mesh of the photographed object in the observation space based on the two-dimensional feature amount information, and change a scale of the mesh based on the result of the comparison between the two-dimensional observation information and the two-dimensional feature amount information.

2. The image processing system according to claim 1 , wherein the at least one processor is configured to partially change the mesh after changing the scale of the mesh based on the result of the comparison between the two-dimensional observation information and the two-dimensional feature amount information.

3. The image processing system according to claim 1 , wherein the additional information includes information on a three-dimensional shape of the photographed object, which is estimated based on the machine learning data.

4. The image processing system according to claim 3 , wherein the additional information includes information on the mesh of the photographed object.

5. The image processing system according to claim 4 , wherein the at least one processor is configured to set the mesh in the observation space based on the additional information, and change the mesh based on the observation space information.

6. The image processing system according to claim 5 , wherein the at least one processor is configured to change a first mesh portion of the mesh that corresponds to the three-dimensional coordinates of the feature point cloud indicated by the observation space information, and then change a second mesh portion around the first mesh portion.

7. The image processing system according to claim 5 ,

wherein the at least one processor is configured to:

estimate a position of the camera based on the changes in position of the feature point cloud, and set an observation viewpoint in the observation space based on a result of the estimation, and

change a first mesh portion based on an orientation of the first mesh portion with respect to the observation viewpoint.

8. The image processing system according to claim 5 , wherein the at least one processor is configured to change scale of the mesh based on the observation space information.

9. The image processing system according to claim 3 , wherein the additional information includes information on a normal of the photographed object.

10. The image processing system according to claim 3 , wherein the additional information includes information on a classification of the photographed object.

11. The image processing system according to claim 1 ,

wherein the camera is configured to take images of the real space based on a predetermined frame rate, and

wherein the at least one processor is configured to execute processing based on one of the taken images that is taken in the same frame.

12. An image processing method, comprising:

a taken image acquisition step of acquiring taken images that have been taken by a camera, which is movable in a real space;

an observation space information acquisition step of acquiring, based on changes in position of a feature point cloud in the taken images, observation space information including three-dimensional coordinates of the feature point cloud in an observation space;

a machine learning step of acquiring, based on machine learning data on a feature of an object, additional information on a feature of a photographed object shown in the taken images; and

an integration step of integrating the observation space information and the additional information;

wherein the additional information includes two-dimensional feature amount information in which a position of the photographed object in the taken images and a feature amount regarding the photographed object are associated with each other,

an estimate step of estimating a position of the camera based on the changes in position of the feature point cloud, and setting an observation viewpoint in the observation space based on a result of the estimation, and

an execute step of executing processing based on a result of comparison between two-dimensional observation information, which shows a view of the observation space as seen from the observation viewpoint, and the two-dimensional feature amount information;

wherein the feature amount includes a depth of the photographed object estimated based on the machine learning data,

wherein, in the two-dimensional observation information, a position of the feature point cloud in a two-dimensional space, and a depth of the feature point cloud in the observation space are associated with each other, and

wherein the at least one processor is configured to set a mesh of the photographed object in the observation space based on the two-dimensional feature amount information, and change a scale of the mesh based on the result of the comparison between the two-dimensional observation information and the two-dimensional feature amount information.

13. A non-transitory computer-readable information storage medium for storing a program for causing a computer to:

acquire taken images that have been taken by a camera, which is movable in a real space;

acquire, based on changes in position of a feature point cloud in the taken images, observation space information including three-dimensional coordinates of the feature point cloud in an observation space;

acquire, based on machine learning data on a feature of an object, additional information on a feature of a photographed object shown in the taken images; and

integrate the observation space information and the additional information;

wherein the additional information includes two-dimensional feature amount information in which a position of the photographed object in the taken images and a feature amount regarding the photographed object are associated with each other,

estimate a position of the camera based on the changes in position of the feature point cloud, and set an observation viewpoint in the observation space based on a result of the estimation, and

execute processing based on a result of comparison between two-dimensional observation information, which shows a view of the observation space as seen from the observation viewpoint, and the two-dimensional feature amount information;

wherein the feature amount includes a depth of the photographed object estimated based on the machine learning data,

wherein, in the two-dimensional observation information, a position of the feature point cloud in a two-dimensional space, and a depth of the feature point cloud in the observation space are associated with each other, and

wherein the at least one processor is configured to set a mesh of the photographed object in the observation space based on the two-dimensional feature amount information, and change a scale of the mesh based on the result of the comparison between the two-dimensional observation information and the two-dimensional feature amount information.