Incremental visual query processing with holistic feature feedback
To reduce communication costs and computational complexity, only a subset of ranked SIFT points within a query image for a visual search request is transmitted to the visual search server in each iteration of an incremental search. For each candidate match, a flag identifying the matching points is returned by the server for use in computing holistic (e.g., histogram) information for a bounding box within the query image including the matching points. Distance from that query image holistic information is used to reject images from a short list used for a subsequent iteration, if any. If all images are rejected or a match criteria is met during one iteration, the search may terminate early without consideration of remaining SIFT points.
1. A method, comprising:
receiving, at a visual search server, information relating to a first subset of distinctive features within a query image for a visual search request;
for at least a subset of stored image information accessible to the visual search server, determining, at a visual search server, one or more sets of stored image information in which information relating to distinctive features for a respective image corresponds to at least a subset of the query image information; and
for each of the one or more sets of stored image information in which information relating to distinctive features for a respective image corresponds to at least a subset of the query image information,
transmitting a flag identifying the corresponding stored image information from the visual search server to a source of the visual search request,
selecting a region from the query image using the flag;
receiving, at the visual search server, holistic information from the visual search request source, the holistic information based upon the selected region, and
using holistic information received at the visual search server from the visual search request source to evaluate a match between the respective image and the query image.
2. The method according to claim 1 , wherein the subset of stored image information comprises a short list of sets of stored image information in which information relating to distinctive features for a respective image corresponded to at least a subset of information relating to a second subset of distinctive features within the query image.
3. The method according to claim 1 , wherein the holistic information comprises one or more of a color histogram for the query image region, a pixel intensity histogram for the query image region, an edge histogram for the query image region, and an appearance model for the query image region.
4. The method according to claim 1 , wherein the information relating to the first subset of distinctive features within the query image comprises one or more of a ranked edge strength and a ranked peak strength.
5. The method according to claim 1 , wherein the first subset of distinctive features within the query image comprise k of n distinctive features, where k<n.
6. A visual search server, comprising:
one or more processors configured to receive, via a network connection, information relating to a first subset of distinctive features within a query image for a visual search request; and
a memory including a plurality of instructions that, when executed by the one or more processors, are configured to cause the one or more processors to:
determine, for at least a subset of stored image information accessible to the visual search server, one or more sets of stored image information in which information relating to distinctive features for a respective image corresponds to at least a subset of the query image information, and
for each of the one or more sets of stored image information in which information relating to distinctive features for a respective image corresponds to at least a subset of the query image information,
transmit a flag identifying the corresponding stored image information from the visual search server to a source of the visual search request,
select a region from the query image using the flag,
receive holistic information from the visual search request source, the holistic information based upon the selected region, and
use holistic information received at the visual search server from the visual search request source evaluate a match between the respective image and the query image.
7. The visual search server according to claim 6 , wherein the subset of stored image information comprises a short list of sets of stored image information in which information relating to distinctive features for a respective image corresponded to at least a subset of information relating to a second subset of distinctive features within the query image.
8. The visual search server according to claim 6 , wherein the holistic information comprises one or more of a color histogram for the query image region, a pixel intensity histogram for the query image region, an edge histogram for the query image region, and an appearance model for the query image region.
9. The visual search server according to claim 6 , wherein the information relating to the first subset of distinctive features within the query image comprises one or more of a ranked edge strength and a ranked peak strength.
10. The visual search server according to claim 6 , wherein the first subset of distinctive features within the query image comprise k of n distinctive features, where k<n.
11. A method, comprising:
transmitting a visual search request containing information relating to a first subset of distinctive features within a query image from a mobile device to a visual search server;
receiving, for each of the one or more sets of stored image information accessible to the visual search server in which information relating to distinctive features for a respective image corresponds to at least a subset of the query image information, a flag identifying the corresponding stored image information from the visual search server;
selecting a region from the query image using the flag;
computing holistic information based upon the selected region; and
transmitting the holistic information to the visual search server for use in evaluating a match between the respective image and the query image.
12. The method according to claim 11 , wherein the subset of stored image information comprises a short list of sets of stored image information in which information relating to distinctive features for a respective image corresponded to at least a subset of information relating to a second subset of distinctive features within the query image.
13. The method according to claim 11 , wherein the holistic information comprises one or more of a color histogram for the query image region, a pixel intensity histogram for the query image region, an edge histogram for the query image region, and an appearance model for the query image region.
14. The method according to claim 11 , wherein the information relating to the first subset of distinctive features within the query image comprises one or more of a ranked edge strength and a ranked peak strength.
15. The method according to claim 11 , wherein the first subset of distinctive features within the query image comprise k of n distinctive features, where k<n.
16. A mobile device, comprising:
a transceiver configured to:
transmit a visual search request containing information relating to a first subset of distinctive features within a query image from a mobile device to a visual search server, and
receive, for each of the one or more sets of stored image information accessible to the visual search server in which information relating to distinctive features for a respective image corresponds to at least a subset of the query image information, a flag identifying the corresponding stored image information from the visual search server;
a processor; and
a memory comprising a plurality of instructions that, when executed by the processor, are configured to cause the processor to:
select a region from the query image using the flag; and
compute holistic information based upon the selected region,
wherein the transceiver is further configured to transmit the holistic information to the visual search server for use in evaluating a match between the respective image and the query image.
17. The mobile device according to claim 16 , wherein the subset of stored image information comprises a short list of sets of stored image information in which information relating to distinctive features for a respective image corresponded to at least a subset of information relating to a second subset of distinctive features within the query image.
18. The mobile device according to claim 16 , wherein the holistic information comprises one or more of a color histogram for the query image region, a pixel intensity histogram for the query image region, an edge histogram for the query image region, and an appearance model for the query image region.
19. The mobile device according to claim 16 , wherein the information relating to the first subset of distinctive features within the query image comprises one or more of a ranked edge strength and a ranked peak strength.
20. The mobile device according to claim 16 , wherein the first subset of distinctive features within the query image comprise k of n distinctive features, where k<n.