Method and apparatus for assigning image location and direction to a floorplan diagram based on artificial intelligence

1. A computer implemented method using artificial intelligence for matching images with locations and directions, the method comprising:

acquiring a plurality of room images taken with a camera;

using artificial intelligence to detect at least one object in each of the plurality of room images;

detecting an object location of the at least one object detected in each of the room images;

acquiring a floorplan image;

using artificial intelligence to detect at least one object in the floorplan image;

detecting an object location of the at least one object detected in the floorplan image;

comparing the at least one object and object location detected in each of the plurality of room images to the at least one object and object location detected in the floorplan image; and

determining a location in the floorplan image where each room image was taken.

2. The method according to claim 1 , wherein a determination is made as to how many degrees each room image must be shifted along a horizontal axis so that each room image is aligned in a first orientation with respect to an orientation in the floorplan image.

3. The method according to claim 1 , wherein a first room image in the plurality of room images is analyzed using a convolutional neural network to determine a classification of each object and the location of each object.

4. The method according to claim 3 , wherein the floorplan image is analyzed using a convolutional neural network to determine the classification of each object and the location of each object in the floorplan image.

5. The method according to claim 1 , further comprising:

performing a distance transform function on the floorplan image to determine the distance between each point in the floorplan image and closest object in the floorplan image.

6. The method according to claim 1 , wherein the detected objects are of a class which includes: door, window, corner or wall.

7. The method according to claim 6 , wherein the comparing the objects and object locations detected in each of the room images to the objects and object locations detected in the floorplan image, is done using a single class of objects at a time.

8. The method according to claim 7 , wherein the comparison between single classes of objects is made by measuring a distance between the objects in the room image and the floorplan image and assigning a cost between objects from a room image and the floorplan image.

9. The method according to claim 8 , wherein a weighting factor is used to determine the cost.

10. The method according to claim 1 , wherein the comparing of the objects and object locations detected in each of the room images to the objects and object locations detected in the floorplan image, is done for each room image and a plurality of locations in the floorplan image.

11. The method according to claim 1 , wherein the comparison between objects is performed using an edit distance.

12. An artificial intelligence (AI) system for matching images with locations and directions, the apparatus comprising a central processing unit (CPU), a graphical processing unit (GPU), and a memory, the CPU configured to:

acquire a plurality of room images;

acquire a floorplan image;

the GPU is configured to:

detect objects and the object locations in each of the room images;

detect objects and the object locations in the floorplan image;

wherein the CPU is further configured to:

compare the objects and object locations detected in each of the room images to the objects and object locations detected in the floorplan image; and

determine a location in the floorplan image where each room image was taken.

13. The AI system according to claim 12 , wherein the AI system is configured to determine how many degrees each room image must be rotated so that each room image is aligned in a first orientation with respect to an orientation in the floorplan image.

14. The AI system according to claim 12 , wherein the GPU implements a convolutional neural network to determine a classification of each object and the location of each object for a first room image in the plurality of room images.

15. The AI system according to claim 14 , wherein the GPU implements a convolutional neural network to determine a classification of each object and the location of each object for the floorplan image.

16. The AI system according to claim 12 , wherein the CPU or GPU is further configured to perform a distance transform function on the floorplan image to determine the distance between each point in the floorplan image and closest object in the floorplan image.

17. The AI system according to claim 12 , wherein the detected objects have classes which include: door, window, corner or wall.

18. The AI system according to claim 17 , wherein the CPU or GPU is further configured to compare the objects and object locations detected in each of the room images to the objects and object locations detected in the floorplan image, is done using a single class of objects at a time.

19. The AI system according to claim 18 , wherein the CPU or GPU is further configured to, when comparing between single classes of objects, determine a distance between the objects in the room image and the floorplan image, and assigning a cost between objects from a room image and the floorplan image.

20. A non-transitory computer readable medium for causing an artificial intelligence (AI) system, which comprises a central processing unit (CPU), a graphical processing unit (GPU), and a memory, to match images with locations of where the images were taken and directions the images were taken in, the non-transitory computer readable medium causing the AI system to:

acquire a plurality of room images;

acquire a floorplan image;

the non-transitory computer readable medium causing the GPU to:

detect objects and the object locations in each of the room images;

detect objects and the object locations in the floorplan image;

wherein the non-transitory computer readable medium further causes the CPU to:

compare the objects and object locations detected in each of the room images to the objects and object locations detected in the floorplan image; and

determine a location in the floorplan image where each room image was taken.