Lazy evaluation of semantic indexing

1. A method for searching a database of digital media assets, comprising:

a processor receiving a search query;

the processor analyzing the search query to extract a specialized search condition corresponding to a specialized semantic concept;

the processor selecting, from a pre-existing ontology of indexers, a specialized indexer based on a determination that the selected specialized indexer can be used to identify digital media assets that correspond to the specialized semantic concept;

the processor analyzing the search query to extract a general search condition corresponding to a related higher level semantic concept;

the processor identifying, from a database of digital media assets, a subset of digital media assets that satisfy the general search condition responsive to general metadata for each of the digital media assets;

the processor applying the selected specialized indexer to the subset of digital media assets to determine specialized metadata for a digital media asset in the subset of digital media assets, wherein the specialized metadata provides an indication of whether the digital media asset corresponds to the specialized semantic concept; and

the processor converting the specialized indexer into a general indexer based at least on a frequency with which the specialized semantic concept is queried, wherein the processor is configured to convert the specialized indexer into the general indexer based further on a computational cost associated with the specialized indexer,

wherein the general indexer is applied to the digital media assets independent of a search query, wherein the general indexer is operable to identify a digital media asset that corresponds to a higher level semantic concept and determine the general metadata for the identified digital media asset based at least upon an analysis of pixel data associated with the digital media asset, and wherein the general metadata provides an indication of the higher level semantic concept to which the identified digital media asset corresponds.

2. The method of claim 1 further comprising ranking the subset of the digital media assets according to their satisfaction of the general search condition responsive to the general metadata.

3. The method of claim 1 wherein an element of the search query corresponds to both a general indexer and a specialized indexer.

4. The method of claim 1 wherein the specialized metadata is associated with the digital media assets and stored in the database of digital media assets.

5. The method of claim 1 wherein the specialized metadata already exists for a portion of the subset of the digital media assets, and wherein the specialized indexer is only applied to digital media assets not having the specialized metadata.

6. The method of claim 1 wherein the search query is received from a user.

7. The method of claim 1 wherein general search conditions are identified from elements of the search query by using concept expansion.

8. The method of claim 1 wherein specialized search conditions are identified from elements of the search query by using concept expansion.

9. The method of claim 6 further comprising:

ranking the subset of the digital media assets according to their satisfaction of the specialized search condition responsive to the specialized metadata; and

presenting the most highly ranked set of digital media assets.

10. The method of claim 9 wherein the digital media assets are digital images and the digital media assets are presented by displaying the digital images.

11. The method of claim 9 wherein the digital media assets are digital video files and the digital media assets are presented by playing the digital video files.

12. The method of claim 9 wherein the digital media assets are digital audio files and the digital media assets are presented by playing the digital audio files.

13. The method of claim 9 wherein the digital media assets are complex objects composed of a plurality of digital images, digital videos or digital audios.

14. The method of claim 1 wherein the general metadata includes date/time metadata, event classification metadata, indoor/outdoor metadata, person detection metadata or geospatial metadata.

15. The method of claim 1 wherein the specialized metadata is a meronym of a generalized metadata term.

16. The method of claim 1 wherein a portion of the general metadata or the specialized metadata are specified by probability values.

17. The method of claim 16 further comprising using the probability values to rank the subset of the digital media assets.

18. The method of claim 1 further comprising presenting a query requesting additional information to clarify the search query.

19. The method of claim 18 wherein the query requests clarification of ambiguous elements of the search query.

20. The method of claim 1 wherein the general indexer is included in a set of general indexers, and wherein the set of general indexers is modified over time in response to a frequency at which search queries are executed containing elements corresponding to each indexer in the set of general indexers.

21. The method of claim 1 , wherein specialized search indexers are applied to digital media assets only in response to particular search queries.

22. The method of claim 1 , further comprising:

ranking the subset of the digital media assets according to their satisfaction of the general search condition based on the general metadata; and

ranking the subset of the digital media assets according to their satisfaction of the specialized search condition based on the specialized metadata.

23. The method of claim 1 , wherein analyzing the search query to extract a general search condition comprises selecting general indexers from the pre-existing ontology, and wherein the pre-existing ontology relates the general indexers to the selected specialized indexers.

24. The method of claim 1 , further comprising selecting a second specialized indexer based on a determination that the pre-existing ontology relates the second specialized indexer to the selected specialized indexer.

25. The method of claim 1 , wherein the pre-existing ontology comprises user-specific ontological relationships.

26. A system comprising:

a data processor; and

a memory communicatively connected to the data processor and storing instructions configured to cause the data processor to perform operations comprising:

receiving a search query;

analyzing the search query to extract a specialized search condition corresponding to a specialized semantic concept;

selecting, from a pre-existing ontology of indexers, a specialized indexer based on a determination that the selected specialized indexer can be used to identify digital media assets that correspond to the specialized semantic concept;

analyzing the search query to extract a general search condition corresponding to a related higher level semantic concept;

identifying, from a database of digital media assets, a subset of digital media assets that satisfy the general search condition responsive to general metadata for each of the digital media assets;

applying the selected specialized indexer to the subset of digital media assets to determine specialized metadata for a digital media asset in the subset of digital media assets, wherein the specialized metadata provides an indication of whether the digital media asset corresponds to the specialized semantic concept; and

the processor converting the specialized indexer into a general indexer based at least on a frequency with which the specialized semantic concept is queried, wherein the processor is configured to convert the specialized indexer into the general indexer based further on a computational cost associated with the specialized indexer;

wherein the general indexer is applied to the digital media assets independent of a search query, wherein the general indexer is operable to identify a digital media asset that corresponds to a higher level semantic concept and determine the general metadata for the identified digital media asset based at least upon an analysis of pixel data associated with the digital media asset, and wherein the general metadata provides an indication of the higher level semantic concept to which the identified digital media asset corresponds.

27. A non-transitory computer readable medium having stored thereon instructions executable by a processor to cause the processor to perform functions comprising:

receiving a search query;

analyzing the search query to extract a specialized search condition corresponding to a specialized semantic concept;

selecting, from a pre-existing ontology of indexers, a specialized indexer based on a determination that the selected specialized indexer can be used to identify digital media assets that correspond to the specialized semantic concept;

analyzing the search query to extract a general search condition corresponding to a related higher level semantic concept;

identifying, from a database of digital media assets, a subset of digital media assets that satisfy the general search condition responsive to general metadata for each of the digital media assets;

applying the selected specialized indexer to the subset of digital media assets to determine specialized metadata for a digital media asset in the subset of digital media assets, wherein the specialized metadata provides an indication of whether the digital media asset corresponds to the specialized semantic concept; and

converting the specialized indexer into a general indexer based at least on a frequency with which the specialized semantic concept is queried,

wherein the processor is configured to convert the specialized indexer into the general indexer based further on a computational cost associated with the specialized indexer,

wherein the general indexer is applied to the digital media assets independent of a search query, wherein the general indexer is operable to identify a digital media asset that corresponds to a higher level semantic concept and determine the general metadata for the identified digital media asset based at least upon an analysis of pixel data associated with the digital media asset, and wherein the general metadata provides an indication of the higher level semantic concept to which the identified digital media asset corresponds.