Back to results
US 12,730,840 B2Granted
Semantic-based font searching in vector space using machine learning
AssigneeSony Interactive Entertainment Inc.
InventorJoseph Logan Olson
Life of this patent
Estimating the term…
Abstract
In one aspect, semantic-based font searching may be performed in vector space using machine learning. Text and images may therefore be put into a same embedded vector space so that font images that are close, in the vector space, to a word from a font search query may be returned as search results. Thus, an end-user can enter a natural language search query string for the system to return fonts that semantically match the user's search string.
Claims20
An apparatus, comprising:
- at least one processor system configured to:
- receive a query representing a semantic-based search for a text font;
- in response to receipt of the query, convert one or more elements of the query into a first vector that represents the semantic-based search within a vector space of fonts where semantically similar fonts are arranged closer to one another than semantically dissimilar fonts;
- search, in the vector space, for vectors representing one or more types of text fonts using the first vector that represents the semantic-based search;
- based on the search, return a second vector that represents a particular text font that semantically matches the query;
- and
- based on the return of the second vector, present, on a graphical user interface (GUI) as a search result to the query, an indication of the particular text font.
4 more elements
A method, comprising:
- receiving a query representing a semantic-based search for a text font;
- in response to receiving the query, converting one or more elements of the query into a first vector that represents the semantic-based search within a vector space of fonts where semantically similar fonts are arranged closer to one another than semantically dissimilar fonts;
- searching, in the vector space, for vectors representing one or more types of text fonts using the first vector that represents the semantic-based search;
- based on the searching, returning a second vector that represents a particular text font that semantically matches the query;
- and
- based on the returning of the second vector, presenting, at a device, an indication of the particular text font.
3 more elements
An apparatus, comprising:
- at least one computer readable storage medium (CRSM) that is not a transitory signal, the at least one CRSM comprising instructions that when executed by a processor system cause the processor system to perform operations comprising:
- receive a query representing a semantic-based search for a text font;
- in response to receipt of the query, convert one or more elements of the query into a first vector that represents the semantic-based search within a vector space of fonts where semantically similar fonts are arranged closer to one another than semantically dissimilar fonts;
- search, in the vector space, for vectors representing one or more types of text fonts using the first vector that represents the semantic-based search;
- based on the search, return a second vector that represents a particular text font that semantically matches the query;
- and
- based on the return of the second vector, present a search result to the query, the search result indicating the particular text font.
4 more elements
Dependencies are inferred from the claim text.