System and method of identifying visual objects
A system and method of identifying objects is provided. In one aspect, the system and method includes a hand-held device with a display, camera and processor. As the camera captures images and displays them on the display, the processor compares the information retrieved in connection with one image with information retrieved in connection with subsequent images. The processor uses the result of such comparison to determine the object that is likely to be of greatest interest to the user. The display simultaneously displays the images the images as they are captured, the location of the object in an image, and information retrieved for the object.
1. A computer-implemented method comprising:
detecting, by a computer system, in a plurality of image frames that include a most recent image frame and a plurality of sequentially prior frames that increase in time prior to the most recent image frame, a plurality of visually distinct features;
obtaining, from the computer systems, a respective description of each of the plurality of features;
for each feature of the plurality of features:
determining matches of respective descriptions associated with the feature in respectively different image frames;
for each respectively different image frame for which a match of the respective description occurred, determining a weight associated with the frame, wherein each weight for each of respective image frame decreases in proportion to the image frames' increase in time prior to the most recent image frame;
determining a score for the feature based on the weights determined for the respectively different image frames for which the match of the respective description occurred;
generating and executing a search query based at least a description having a highest determined score relative to the determined scores of other descriptions.
2. The method of claim 1 , wherein detecting the plurality of visually distinct features comprises detecting object depicted in the image frames.
3. The method of claim 1 , wherein obtaining a respective description of each of the plurality of features comprises obtaining, for each feature, text that describes the feature.
4. The method of claim 3 , wherein determining matches of respective descriptions associated with the feature in respectively different image frames comprises determining that the text of a feature for a first image frame matches the text of a feature for a second image frame.
5. The method of claim 1 , wherein generating and executing a search query based at least a description having a highest determined score relative to the determined scores of other descriptions comprises generating a search query that includes one or more search terms that correspond to descriptions.
6. The method of claim 1 , wherein determining a weight associated with the frame, wherein each weight for each of respective image frame decreases in proportion to the image frames' increase in time prior to the most recent image frame comprises:
determining a weight for the most recent image frame;
for each successively prior image from beginning from the image frame immediately prior to the most recent frame that occurs, determining a weight that is a fraction of the weight determined for a frame that occurs immediately after the prior image frame.
7. The method of claim 1 , wherein determining a score for the feature based on the weights determined for the respectively different image frames for which the match of the respective description occurred comprise aggregating the weights determined for the respectively different image frames for which the match of the respective description occurred.
8. A system comprising:
one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
detecting in a plurality of image frames that include a most recent image frame and a plurality of sequentially prior frames that increase in time prior to the most recent image frame, a plurality of visually distinct features;
obtaining a respective description of each of the plurality of features;
for each feature of the plurality of features:
determining matches of respective descriptions associated with the feature in respectively different image frames;
for each respectively different image frame for which a match of the respective description occurred, determining a weight associated with the frame, wherein each weight for each of respective image frame decreases in proportion to the image frames' increase in time prior to the most recent image frame;
determining a score for the feature based on the weights determined for the respectively different image frames for which the match of the respective description occurred;
generating and executing a search query based at least a description having a highest determined score relative to the determined scores of other descriptions.
9. The system of claim 8 , wherein detecting the plurality of visually distinct features comprises detecting object depicted in the image frames.
10. The system of claim 8 , wherein obtaining a respective description of each of the plurality of features comprises obtaining, for each feature, text that describes the feature.
11. The system of claim 10 , wherein determining matches of respective descriptions associated with the feature in respectively different image frames comprises determining that the text of a feature for a first image frame matches the text of a feature for a second image frame.
12. The system of claim 8 , wherein generating and executing a search query based at least a description having a highest determined score relative to the determined scores of other descriptions comprises generating a search query that includes one or more search terms that correspond to descriptions.
13. The system of claim 8 , wherein determining a weight associated with the frame, wherein each weight for each of respective image frame decreases in proportion to the image frames' increase in time prior to the most recent image frame comprises:
determining a weight for the most recent image frame;
for each successively prior image from beginning from the image frame immediately prior to the most recent frame that occurs, determining a weight that is a fraction of the weight determined for a frame that occurs immediately after the prior image frame.
14. The system of claim 8 , wherein determining a score for the feature based on the weights determined for the respectively different image frames for which the match of the respective description occurred comprise aggregating the weights determined for the respectively different image frames for which the match of the respective description occurred.
15. A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:
detecting, in a plurality of visually distinct features including a first feature and a second feature;
obtaining a respective description of each of the plurality of features, including a first description associated with the first feature and a second description associated with the second feature;
determining that the first description associated with the first feature matches the second description associated with the second feature;
determining that a first frequency count of the first description matching other descriptions in the plurality of features exceeds a frequency count of any other description in the descriptions of the plurality of features;
generating and executing a search query based at least on the first description and the first frequency count; and
selecting a highest ranking result from results returned by the search query as an optimum annotation for a first image in the plurality of images.
16. The non-transitory computer-readable medium of claim 15 , wherein determining a weight associated with the frame, wherein each weight for each of respective image frame decreases in proportion to the image frames' increase in time prior to the most recent image frame comprises:
determining a weight for the most recent image frame;
for each successively prior image from beginning from the image frame immediately prior to the most recent frame that occurs, determining a weight that is a fraction of the weight determined for a frame that occurs immediately after the prior image frame.
17. The non-transitory computer-readable medium of claim 15 , wherein determining a score for the feature based on the weights determined for the respectively different image frames for which the match of the respective description occurred comprise aggregating the weights determined for the respectively different image frames for which the match of the respective description occurred.