Automated image retrieval with image graph
An image retrieval system receives an image for which to identify relevant images from an image repository. Relevant images may be of the same environment or object and features and other characteristics. Images in the repository are represented in an image retrieval graph by a set of image nodes connected by edges to other related image nodes with edge weights representing the similarity of the nodes to each other. Based on the received image, the image traversal system identifies an image in the image retrieval graph and alternatively explores and traverses (also termed “exploits”) the image nodes with the edge weights. In the exploration step, image nodes in an exploration set are evaluated to identify connected nodes that are added to a traversal set of image nodes. In the traversal step, the relevant nodes in the traversal set are added to the exploration set and a query result set.
1. A computer system for automated object retrieval, the computer system comprising:
a processor configured to execute instructions; and
a non-transient computer-readable medium comprising instructions that when executed by the processor cause the processor to:
identify a query object represented in a retrieval graph having object nodes representing objects connected by a set of edges having edge weights representing similarity between the objects represented by the connected object nodes; and
determine a query result set of relevant objects to the query object by iteratively:
identifying a relevant object node in a traversal set of object nodes based on an associated traversal weight of the object node;
adding the relevant object node to the query result set; and
updating the traversal set with one or more object nodes connected to the relevant object node with associated traversal weights based on the edge weights between the relevant object node and the respective object node.
2. The system of claim 1 , wherein the objects are images.
3. The system of claim 1 , wherein the instructions further cause the processor to:
receive a request from a client device to retrieve objects relevant to the query object; and
return the query result set to the client device.
4. The computer system of claim 1 , wherein the retrieval graph is sparsely connected.
5. The computer system of claim 1 , wherein a set of relevant object nodes is identified and added to the query result set as an ordered queue based on respective traversal weights of the relevant object nodes in the traversal set.
6. The computer system of claim 1 , wherein the traversal set of object nodes is a heap prioritized by traversal weights.
7. The computer system of claim 1 , wherein the object nodes connected in the retrieval graph are selected based on a ranked number of a k-nearest neighbor (k-NN).
8. The computer system of claim 1 , wherein the edge weights of the set of edges in the retrieval graph is based on an inlier score of a Random Sample Consensus (RANSAC) algorithm.
9. A method for automated image retrieval, the method comprising:
identifying a query object represented in a retrieval graph having object nodes representing objects connected by a set of edges having edge weights representing similarity between the objects represented by the connected object nodes;
determining a query result set of relevant objects to the query object by iteratively:
identifying a relevant object node in a traversal set of object nodes based on an associated traversal weight of the object node;
adding the relevant object node to the query result set; and
updating the traversal set with one or more object nodes connected to the relevant object node with associated traversal weights based on the edge weights between the relevant object node and the respective object node.
10. The method of claim 9 , wherein the objects are images.
11. The method of claim 9 , further comprising:
receiving a request from a client device to retrieve objects relevant to the query object; and
returning the query result set to the client device.
12. The method of claim 9 , wherein the retrieval graph is sparsely connected.
13. The method of claim 9 , wherein a set of relevant object nodes is identified and added to the query result set as an ordered queue based on respective traversal weights of the relevant object nodes in the traversal set.
14. The method of claim 9 , wherein the traversal set of object nodes is a heap prioritized by traversal weights.
15. The method of claim 9 , wherein the object nodes connected in the image retrieval graph are selected based on a ranked number of a k-nearest neighbor (k-NN).
16. The method of claim 9 , wherein the edge weights of the set of edges in the image retrieval graph is based on an inlier score of a Random Sample Consensus (RANSAC) algorithm.
17. A non-transitory computer-readable medium containing computer program code that is executable by a processor for the processor to perform steps of:
identifying a query object represented in a retrieval graph having object nodes representing objects connected by a set of edges having edge weights representing similarity between the objects represented by the connected object nodes; and
determining a query result set of relevant objects to the query object by iteratively:
identifying a relevant object node in a traversal set of object nodes based on an associated traversal weight of the object node;
adding the relevant object node to the query result set; and
updating the traversal set with one or more object nodes connected to the relevant object node with associated traversal weights based on the edge weights between the relevant object node and the respective object node.
18. The non-transitory computer-readable medium of claim 17 , wherein the traversal set of object nodes is a heap prioritized by traversal weights.
19. The non-transitory computer-readable medium of claim 17 , wherein the object nodes connected in the retrieval graph are selected based on a ranked number of a k-nearest neighbor (k-NN).
20. The non-transitory computer-readable medium of claim 17 , wherein the edge weights of the set of edges in the retrieval graph is based on an inlier score of a Random Sample Consensus (RANSAC) algorithm.