Feature selection for image based location determination

1. A computer-implemented method comprising:

receiving, by one or more computing devices having one or more processors, information collected at a particular indoor area;

identifying, by the one or more computing devices, a plurality of objects of the particular indoor area and associated locations from the received information;

for each given of the identified plurality of objects, generating, by the one or more computing devices, a value indicative of a uniqueness in a geometric shape of the given object in the particular indoor area based on a total number of other objects of the plurality of objects that are of a same type of object as the given object, wherein generated values for objects that (1) do not appear in other indoor areas and (2) are not moved to a new location indicate greater uniqueness than generated values for objects that either:

(1) appear in other indoor areas, or

(2) are moved to a new location;

filtering, by the one or more computing devices, the identified plurality of objects to remove any objects associated with generated values indicative of objects that appear in other indoor areas and to remove any objects associated with generated values indicative of objects that are moved to a new location;

ranking, by the one or more computing devices, the filtered identified plurality of objects based on the generated values for the filtered identified plurality of objects; and

generating, by the one or more computing devices, a set of geometry data for locating a mobile device at the particular area based on the ranking and the associated locations such that the set of geometry data includes data corresponding to a most unique object of the plurality of objects,

wherein the set of geometry data includes a plurality of points and vectors defining the associated locations and shapes of the plurality of objects including the most unique object of the plurality of objects within the particular indoor area.

2. The method of claim 1 , wherein the set of geometry data is generated further based on a threshold ranking percentile.

3. The method of claim 1 , wherein the set of geometry data is generated further based on a number of directions a camera of a mobile computing device is likely to be pointed in from a given location at the particular area.

4. The method of claim 1 , wherein the set of geometry data is generated such that the set of geometry data does not include duplicate objects.

5. The method of claim 1 , wherein the set of geometry data is generated further based on a size of the set of geometry data when the geometry data is compressed.

6. The method of claim 1 , wherein the set of geometry data is generated further based on expected locations of people at the particular area.

7. The method of claim 1 , further comprising sending the set of geometry data to a client computing device.

8. A system comprising one or more computing devices, having one or more processors, the one or more computing devices being configured to:

receive information collected at a particular indoor area;

identify a plurality of objects of the particular indoor area and associated locations from the received information;

for each given of the identified plurality of objects, generate a value indicative of a uniqueness in a geometric shape of the given objects in the particular indoor area based on a total number of other objects of the plurality of objects that are of a same type of object as the given object, wherein generated values for objects that (1) do not appear in other indoor areas and (2) are not moved to a new location appear to indicate greater uniqueness than generated values for objects that either:

(1) appear in other indoor areas, or

(2) are moved to a new location;

filter the identified plurality of objects to remove any objects associated with generated values indicative of objects that appear in other indoor areas and to remove any objects associated with generated values indicative of objects that are moved to a new location;

rank the filtered identified plurality of objects based on the generated values for the filtered identified plurality of objects; and

generate a set of geometry data for locating a mobile device at the particular area based on the ranking and the associated locations such that the set of geometry data includes data corresponding to a most unique object of the plurality of objects,

wherein the set of geometry data includes a plurality of points and vectors defining the associated locations and shapes of the plurality of objects including the most unique object of the plurality of objects within the particular indoor area.

9. The system of claim 8 , wherein the one or more computing devices are further configured to generate the set of geometry data further based on a threshold ranking percentile.

10. The system of claim 8 , wherein the one or more computing devices are further configured to generate the set of geometry data further based on a number of directions a camera of a mobile computing device is likely to be pointed in from a given location at the particular area.

11. The system of claim 8 , wherein the one or more computing devices are further configured to generate the set of geometry data such that the set of geometry data does not include duplicate objects.

12. The system of claim 8 , wherein the one or more computing devices are further configured to generate the set of geometry data further based on a size of the set of geometry data when the geometry data is compressed.

13. The system of claim 8 , wherein the one or more computing devices are further configured to generate the set of geometry data further based on expected locations of people at the particular area.

14. The system of claim 8 , wherein the one or more computing devices are further configured to send the set of geometry data to a client computing device.

15. A non-transitory, computer-readable storage medium on which computer readable instructions of a program are stored, the instructions, when executed by one or more processors, cause the one or more processors to perform a method, the method comprising:

receiving information collected at a particular indoor area;

identifying a plurality of objects of the particular indoor area and associated locations from the received information;

for each given of the identified plurality of objects, generating a value indicative of a uniqueness in a geometric shape of the given object in the particular indoor area based on a total number of other objects of the plurality of objects that are of a same type of object as the given object, wherein generated values for objects that (1) do not appear in other indoor areas and (2) are not moved to a new location indicate greater uniqueness than generated values for objects that either:

(1) appear in other indoor areas, or

(2) are moved to a new location;

filtering, by the one or more computing devices, the identified plurality of objects to remove any objects associated with generated values indicative of objects that appear in other indoor areas and to remove any objects associated with generated values indicative of objects that are moved to a new location;

ranking the filtered identified plurality of objects based on the generated values for the filtered identified plurality of objects; and

generating a set of geometry data for locating a mobile device at the particular area based on the ranking and the associated locations such that the set of geometry data includes data corresponding to a most unique object of the plurality of objects,

wherein the set of geometry data includes a plurality of points and vectors defining the associated locations and shapes of the plurality of objects including the most unique object of the plurality of objects within the particular indoor area.

16. The medium of claim 15 , wherein the set of geometry data is generated further based on a threshold ranking percentile.

17. The medium of claim 15 , wherein the set of geometry data is generated further based on a size of the set of geometry data when the geometry data is compressed.

18. The medium of claim 15 , wherein the set of geometry data is generated further based on expected locations of people at the particular area.

19. The medium of claim 15 , wherein the set of geometry data is generated further based on a number of directions a camera of a mobile computing device is likely to be pointed in from a given location at the particular area.

20. The medium of claim 15 , wherein the set of geometry data is generated such that the set of geometry data does not include duplicate objects.