Identifying reliable and authoritative sources of multimedia content
Embodiments are directed towards identifying authoritative sources of multimedia content useable in rank ordering class-dependent search-query results of multimedia content. In one embodiment, a citation model is employed for measuring or otherwise determining a strength of an authority to a content source. In one embodiment, a directed graph is constructed over a network of sources based on a propensity of one source to “cite” content provided by another source. In one embodiment, a random walk may be conducted across the network of sources to arrive at authority scores for each source in the network. In another embodiment, a machine-learning algorithm may be used to arrive at authority scores. The authority scores may then be applied for ranking, for example, search-query results, and/or retrieval purposes.
1. A network device, comprising:
a transceiver to send and receive data over a network; and
a processor that operates on the received data to perform actions, including:
receiving a plurality of multimedia content including information about the content for each of the plurality of multimedia content;
employing the information about the content to identify a source for each of the multimedia content;
employing a weighted citation network of relationships based on the identified source for each multimedia content to identify authority scores for each source of the multimedia contents, the citation of network of relationships identifies a citation of multimedia content by network locations, and where the citation of the multimedia content is determined based on detecting that the multimedia content is copied instead of linked from one network location to an other network location;
receiving a class-dependent search-query for other multimedia content; and
employing the identified authority scores to rank order a display of a search-query result to the class-dependent search-query.
2. The network device of claim 1 , wherein the multimedia content comprises images and near-duplicate detection is employed to identify related multimedia content that is used to generate the citation network of relationships.
3. The network device of claim 2 , wherein the near-duplicate detection further comprises employing an ambiguity rejection mechanism, in part by, measuring a distance between scale-invariant feature transform descriptors within at least two multimedia content images applying a threshold to determine whether the two multimedia content images match bi-directionally.
4. The network device of claim 1 , wherein the information about the content comprises at least one of a “last-modified” date or image modification information, and wherein the “last-modified” date or image modification information is used to identify a source ordering of multimedia content within the citation network of relationships.
5. The network device of claim 1 , wherein the processor operates to perform in actions, further comprising:
performing a plurality of class-dependent search queries for multimedia content over a social image sharing site; and
wherein the information about the content includes whether a multimedia content is identified by a user within the social image sharing site as a favorite.
6. The network device of claim 1 , wherein determining an authority score for each source within the citation network further comprises employing a machine-learning algorithm.
7. A system, comprising:
one or more network devices, each having stored thereon a plurality of multimedia content; and
one or more other network devices, each other network device operates to access at least some of the plurality of multimedia content, and to perform actions, comprising:
performing a plurality of class-dependent search queries for multimedia content;
receiving multimedia content results for each of the plurality of queries including content information about each of multimedia content results;
employing the content information to identify a source for each of the multimedia content result;
employing a weighted citation network of relationships based on the identified source for each multimedia content to identify authority scores for each source within the weighted citation network, where the citation of the multimedia content is determined based on detecting that the multimedia content is copied instead of linked from one network location to an other network location;
receiving a class-dependent search-query for multimedia content; and
employing the identified authority scores to rank order a display of a search-query result.
8. The system of claim 7 , wherein the weighted citation network of relationships is generated to identify a citation of multimedia content within a query result.
9. The system of claim 7 , wherein the multimedia content comprises images and near-duplicate detection is employed to identify related multimedia content.
10. The system of claim 9 , wherein the near-duplicate detection further comprises employing an ambiguity rejection mechanism, in part by, measuring a distance between scale-invariant feature transform descriptors within at least two multimedia content images applying a threshold to determine whether the two multimedia content images match bi-directionally.
11. The system of claim 7 , wherein the content information comprises at least one of a “last-modified” date or image modification information, and wherein the “last-modified” date or image modification information is used to identify a source ordering of multimedia content within the citation network.
12. The system of claim 7 , wherein the plurality of class-dependent search queries are performed over a social image sharing site, and the content information includes whether an multimedia content is identified by a user within the social image sharing site as a favorite.
13. A non-transitory computer-readable storage device having computer-executable instructions, the computer-executable instructions when installed onto a computing device enable the computing device to perform actions, comprising:
performing a plurality of class-dependent search queries for multimedia content;
receiving multimedia content results for each of the plurality of queries including content information about each of multimedia content results;
employing the content information to identify a source for each of the multimedia content result;
employing a weighted citation network of relationships based on the identified source for each multimedia content to identify authority scores for each source within the weighted citation network, the citation of network of relationships identifies a citation of multimedia content by network locations, and where the citation of the multimedia content is determined based on detecting that the multimedia content is copied instead of linked from one network location to an other network location;
receiving a class-dependent search-query for multimedia content; and employing the identified authority scores to rank order a display of a search-query result.
14. The non-transitory computer-readable storage device of claim 13 , wherein the weighted citation network of relationships identifies a citation of multimedia content within a query result.
15. The non-transitory computer-readable storage device of claim 13 , wherein the multimedia content comprises images and near-duplicate detection is employed to identify related multimedia content that is used to generate the citation network of relationships.
16. The non-transitory computer-readable storage device of claim 15 , wherein the near-duplicate detection further comprises employing an ambiguity rejection mechanism, in part by, measuring a distance between scale-invariant feature transform descriptors within at least two multimedia content images applying a threshold to determine whether the two multimedia content images match bi-directionally.
17. The non-transitory computer-readable storage device of claim 13 , wherein the content information comprises at least one of a “last-modified” date or image modification information, and wherein the “last-modified” date or image modification information is used to identify a source ordering of multimedia content within the citation network.
18. The non-transitory computer-readable storage device of claim 13 , wherein the plurality of class-dependent search queries are performed over a social image sharing site, and the content information includes whether an multimedia content is identified by a user within the social image sharing site as a favorite.
19. The non-transitory computer-readable storage device of claim 14 , wherein determining an authority score for each source within the citation network further comprises employing a machine-learning algorithm.