Exploration of large-scale data sets

1. A method for image exploration, comprising:

identifying a set of images;

reducing the set of images to obtain a representative set of images that is distributed throughout the set of images by removing a neighbor image based on a proximity of the neighbor image to an image of the representative set of images;

arranging the representative set of images in a grid structure using a self-sorting map (SSM) algorithm; and

displaying a portion of the representative set of images based on the grid structure.

2. The method of claim 1 , further comprising:

receiving a navigation input corresponding to an image of the representative set of images;

identifying a set of neighbor images of the image from among the set of images, wherein at least one image in the set of neighbor images is not included in the representative set of images;

arranging the set of neighbor images in a second grid structure based on the navigation input; and

displaying the set of neighbor images based on the second grid structure.

3. The method of claim 2 , further comprising:

embedding the set of images to obtain an embedded set of images;

adding the embedded set of images to a nearest neighbor index; and

identifying the set of neighbor images based on the nearest neighbor index.

4. The method of claim 1 , further comprising:

identifying a plurality of subsets of the representative set of images; and

selecting a representative image from each of the plurality of subsets of the representative set of images.

5. The method of claim 4 , further comprising:

identifying a centroid for each of the plurality of subsets of the representative set of images; and

selecting the representative image based on the centroid.

6. The method of claim 4 , further comprising:

including the representative image from each of the plurality of subsets of the representative set of images in a second grid structure;

receiving a navigation input; and

displaying the representative images based on the second grid structure in response to the navigation input.

7. The method of claim 1 , further comprising:

generating a graph of the set of images; and

removing the neighbor image based on the graph.

8. The method of claim 7 , further comprising:

generating an image embedding for each of the set of images, wherein the graph is based on the embedding.

9. The method of claim 7 , further comprising:

computing a length of an edge of the graph between the image and the neighbor image, wherein the proximity is based on the length of the edge.

10. The method of claim 1 , wherein:

the proximity is determined according to a self-organizing map algorithm.

11. The method of claim 1 , further comprising:

computing the proximity of the image to the neighbor image using a nearest neighbor algorithm.

12. The method of claim 1 , wherein:

the representative set of images is evenly distributed across the set of images.

13. The method of claim 1 , further comprising:

identifying a first edge of the grid structure and a second edge of the grid structure opposite to the first edge in a first direction; and

connecting the first edge and the second edge to enable continuous navigation around the grid structure in the first direction.

14. The method of claim 13 , further comprising:

identifying a third edge of the grid structure and a fourth edge of the grid structure opposite to the third edge in a second direction; and

connecting the third edge and the fourth edge to enable continuous navigation around the grid structure in the second direction.

15. The method of claim 1 , further comprising:

determining a location identifier, wherein the portion of the representative set of images is displayed based on the location identifier.

16. The method of claim 15 , further comprising:

assigning a bookmark to the location identifier at a first time; and

displaying the portion of the representative set of images at a second time after the first time based on the bookmark.

17. The method of claim 15 , further comprising:

generating a minimap for the representative set of images; and

displaying a navigation position of the location identifier in the minimap.

18. A method for image exploration, comprising:

identifying a set of images;

reducing the set of images to obtain a representative set of images;

arranging the representative set of images in a grid structure using a self-sorting map (SSM) algorithm;

displaying a portion of the representative set of images based on the grid structure;

receiving a navigation input corresponding to an image of the representative set of images;

identifying a set of neighbor images of the image from among the set of images, wherein at least one image in the set of neighbor images is not included in the representative set of images;

arranging the set of neighbor images in a second grid structure based on the navigation input; and

displaying the set of neighbor images based on the second grid structure.

19. The method of claim 18 , further comprising:

removing a neighbor image based on a proximity of the neighbor image to an image of the representative set of images, wherein the representative set of images is distributed throughout the set of images.

20. An apparatus for image exploration, comprising:

a processor;

a memory storing instructions executable by the processor;

a reduction component configured to identify a set of images and to reduce the set of images to obtain a representative set of images by removing a neighbor image based on a proximity of the neighbor image to an image of the representative set of images;

a sorting component configured to arrange the representative set of images in a grid structure using a self-sorting map (SSM) algorithm; and

a user interface configured to display a portion of the representative set of images based on the grid structure and to receive a navigation input, wherein the portion of the representative set of images is displayed based on the navigation input.