Associating still images and videos
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for associating still images and videos. One method includes receiving a plurality of images and a plurality of videos and determining whether the images are related to the videos. The determining includes, for an image and a video, extracting features from the image and extracting features frames of the video, and comparing the features to determine whether the image is related to the video. The method further includes maintaining a data store storing data associating each image with each video determined to be related to the image.
1. A system comprising one or more computers including one or more storage devices storing instructions that when executed by the one or more computers cause the one or more computers to perform operations comprising:
receiving a search query submitted by a user of a user device;
identifying one or more images that are responsive to the search query;
determining that a video is related to one or more of the images responsive to the search query, wherein determining that the video is related to a particular image responsive to the search query comprises:
extracting one or more features from the particular image responsive to the search query;
extracting one or more features from the video;
comparing the features extracted from the particular image responsive to the search query to the features extracted from the video; and
determining that the video is related to the particular image responsive to the search query based on the comparison of the features extracted from the particular image responsive to the search query to the features extracted from the video; and
presenting the video alongside the images that are responsive to the search query on a display of the user device.
2. The system of claim 1 , wherein extracting one or more features from the video comprises:
identifying one or more representative frames in the video, wherein each representative frame is representative of a portion of the video; and
extracting one or more respective features from each representative frame.
3. The system of claim 1 , wherein the images are still images.
4. The system of claim 1 , wherein determining that the video is related to the particular image responsive to the search query based on the comparison of the features extracted from the particular image responsive to the search query to the features extracted from the video comprises:
determining a strength of relationship between the particular image and the video based on the comparison of the features extracted from the particular image to the features extracted from the video; and
determining that the video is related to the particular image if the strength of relationship satisfies a threshold.
5. The system of claim 4 , wherein the strength of relationship between the particular image and the video is an estimate of visual similarity between the particular image and the video.
6. The system of claim 1 , further comprising determining a category of the video and a category of the particular image;
wherein extracting features from the video comprises extracting features based on the category of the video; and
wherein extracting features from the particular image comprises extracting features based on the category of the particular image.
7. The system of claim 6 , wherein determining the category of the video comprises determining the category from text associated with the video, and determining the category of the particular image comprises determining the category from text associated with the particular image.
8. The system of claim 1 , wherein the video is presented alongside the images that are responsive to the search query on the display of the user device in response to the search query submitted by the user and prior to the user submitting subsequent search queries.
9. A method performed by one or more data processing apparatus, the method comprising:
receiving a search query submitted by a user of a user device;
identifying one or more images that are responsive to the search query;
determining that a video is related to one or more of the images responsive to the search query, wherein determining that the video is related to a particular image responsive to the search query comprises:
extracting one or more features from the particular image responsive to the search query;
extracting one or more features from the video;
comparing the features extracted from the particular image responsive to the search query to the features extracted from the video; and
determining that the video is related to the particular image responsive to the search query based on the comparison of the features extracted from the particular image responsive to the search query to the features extracted from the video; and
presenting the video alongside the images that are responsive to the search query on a display of the user device.
10. The method of claim 9 , wherein extracting one or more features from the video comprises:
identifying one or more representative frames in the video, wherein each representative frame is representative of a portion of the video; and
extracting one or more respective features from each representative frame.
11. The method of claim 9 , wherein the images are still images.
12. The method of claim 9 , wherein determining that the video is related to the particular image responsive to the search query based on the comparison of the features extracted from the particular image responsive to the search query to the features extracted from the video comprises:
determining a strength of relationship between the particular image and the video based on the comparison of the features extracted from the particular image to the features extracted from the video; and
determining that the video is related to the particular image if the strength of relationship satisfies a threshold.
13. The method of claim 12 , wherein the strength of relationship between the particular image and the video is an estimate of visual similarity between the particular image and the video.
14. The method of claim 9 , further comprising determining a category of the video and a category of the particular image;
wherein extracting features from the video comprises extracting features based on the category of the video; and
wherein extracting features from the particular image comprises extracting features based on the category of the particular image.
15. The method of claim 14 , wherein determining the category of the video comprises determining the category from text associated with the video, and determining the category of the particular image comprises determining the category from text associated with the particular image.
16. One or more non-transitory computer storage media storing instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:
receiving a search query submitted by a user of a user device;
identifying one or more images that are responsive to the search query;
determining that a video is related to one or more of the images responsive to the search query, wherein determining that the video is related to a particular image responsive to the search query comprises:
extracting one or more features from the particular image responsive to the search query;
extracting one or more features from the video;
comparing the features extracted from the particular image responsive to the search query to the features extracted from the video; and
determining that the video is related to the particular image responsive to the search query based on the comparison of the features extracted from the particular image responsive to the search query to the features extracted from the video; and
presenting the video alongside the images that are responsive to the search query on a display of the user device.
17. The non-transitory computer storage media of claim 16 , wherein extracting one or more features from the video comprises:
identifying one or more representative frames in the video, wherein each representative frame is representative of a portion of the video; and
extracting one or more respective features from each representative frame.
18. The non-transitory computer storage media of claim 16 , wherein the images are still images.
19. The non-transitory computer storage media of claim 16 , wherein determining that the video is related to the particular image responsive to the search query based on the comparison of the features extracted from the particular image responsive to the search query to the features extracted from the video comprises:
determining a strength of relationship between the particular image and the video based on the comparison of the features extracted from the particular image to the features extracted from the video; and
determining that the video is related to the particular image if the strength of relationship satisfies a threshold.
20. The non-transitory computer storage media of claim 19 , wherein the strength of relationship between the particular image and the video is an estimate of visual similarity between the particular image and the video.