System and method for improving speed of similarity based searches
A method and system for processing images for a search is provided, including: receiving a plurality of images selected from search results; for each image in the plurality of images, retrieving a feature vector associated with the image; selecting a subset of the feature vectors based on similarity of feature vectors associated with the images in the plurality of images; and performing a search for feature vectors in a database similar to the feature vectors in the subset of feature vectors.
1. A method of processing images for a search, comprising:
receiving a plurality of images selected from search results;
for each image in the plurality of images, retrieving a respective feature vector associated therewith;
selecting a subset of the feature vectors based on similarity of the feature vectors; and
performing a search for feature vectors in a database similar to feature vectors in the subset of feature vectors, and
wherein the plurality of images come from video captured by a camera.
2. The method of claim 1 wherein the received plurality of images are selected based on similarity to a reference image.
3. The method of claim 2 wherein the plurality of images are selected by a user.
4. The method of claim 1 wherein the plurality of images were generated from a search for images similar to a reference image.
5. The method of claim 1 wherein selecting the subset of feature vectors comprises clustering the feature vectors associated with the images into a plurality of clusters.
6. The method of claim 5 wherein selecting the subset of images comprises filtering the feature vectors based on the clustering.
7. The method of claim 6 wherein the clustering is k-mediod clustering.
8. The method of claim 7 wherein the filtering is selecting a feature vector from each of the clusters.
9. The method of claim 8 wherein selecting a feature vector from each of the clusters comprises selecting a feature vector in a cluster associated with an image showing the face of a person.
10. The method of claim 1 wherein the number of images in the plurality of images exceeds a threshold.
11. The method of claim 10 wherein the number of feature vectors in the subset of feature vectors does not exceed the threshold.
12. The method of claim 1 further comprising displaying the images associated with the feature vectors located in the search.
13. A method of processing images for a search, comprising:
conducting a search for images similar to a reference image;
receiving a plurality of images selected from search results;
for each image in the plurality of images, retrieving a respective feature vector associated therewith;
selecting a subset of the feature vectors based on similarity of the feature vectors;
locating feature vectors in a database similar to feature vectors in the subset of feature vectors; and
displaying images associated with the located feature vectors.
14. The method of claim 13 wherein selecting the subset of feature vectors comprises clustering the feature vectors associated with the images into a plurality of clusters.
15. The method of claim 14 wherein selecting the subset of images comprises filtering the feature vectors based on the clustering.
16. The method of claim 15 wherein the clustering is k-mediod clustering.
17. The method of claim 16 wherein the filtering is selecting a feature vector from each of the clusters.
18. The method of claim 17 wherein selecting a feature vector from each of the clusters comprises selecting a feature vector in a cluster associated with an image showing the face of a person.
19. The method of claim 13 wherein the number of images in the plurality of images exceeds a threshold.