Content collection search with robust content matching

A computer-implemented method for searching a content collection, comprising:

• under control of one or more computer systems configured with executable instructions,
• identifying query descriptors for query content of a search request, the query descriptors characterizing the query content;
• for each content piece of at least a portion of a content collection:
• identifying a subset of content descriptors of the content piece that corresponds to at least a portion of the query descriptors, the content descriptors characterizing the content piece, one or more of the subset of the content descriptors corresponding to one or more blur transforms;
• identifying first candidate regions of the content piece, one or more of the first candidate regions overlapping with at least another one of the first candidate regions, each first candidate region corresponding to at least a portion of the subset of the content descriptors;
• and
• selecting the content piece for inclusion in a matching content subset of the content collection when at least one of the first candidate regions comprises a first proportionate size greater than a first proportion threshold;
• and
• providing the matching content subset in response to the search request.

A non-transitory computer readable storage medium storing instructions for searching a content collection, the instructions when executed by a processor causing the processor to:

• identify query descriptors for query content of a search request, the query descriptors characterizing the query content;
• for each content piece of at least a portion of a content collection:
• identify a subset of content descriptors of the content piece that corresponds to at least a portion of the query descriptors, the content descriptors characterizing the content piece, one or more of the subset of the content descriptors corresponding to one or more blur transforms;
• identify first candidate regions of the content piece, one or more of the first candidate regions overlapping with at least another one of the first candidate regions, each first candidate region corresponding to at east a portion of the subset of the content descriptors;
• and
• select the content piece for inclusion in a matching content subset of the content collection when at least one of the first candidate regions comprises a first proportionate size greater than a first proportion threshold;
• and
• provide the matching content subset in response to the search request.

A system, comprising:

• a processor;
• memory including instructions that, upon being executed by the processor, cause the system to:
• identify query descriptors for query content of a search request, the query descriptors characterizing the query content;
• for each content piece of at least a portion of a content collection:
• identify a subset of content descriptors of the content piece that corresponds to at least a portion of the query descriptors, the content descriptors characterizing the content piece, one or more of the subset of the content descriptors corresponding to one or more blur transforms;
• identify first candidate regions of the content piece, one or more of the first candidate regions overlapping with at least another one of the first candidate regions, each first candidate region corresponding to at least a portion of the subset of the content descriptors;
• and
• select the content piece for inclusion in a matching content subset of the content collection when at least one of the first candidate regions comprises a first proportionate size greater than a first proportion threshold;
• and
• provide the matching content subset in response to the search request.