Linguistic analysis and correction
AssigneeDisney Enterprises, Inc.
InventorsTadashi George Sakashita, Imran Goychayev, Alejandra Ledesma, Luigi Pardo, Sonia Souto Sieiro
Life of this patent
Estimating the term…
Abstract
Methods, computer program products, and systems for correcting a glyph in a translated text are described. In one embodiment, the method includes identifying a first form of a first glyph in a translation text having a plurality of contextual properties and analyzing, by the processor, the first form of the first glyph with reference to one or more glyph form tables comprising a plurality of forms of the first glyph based, at least in part, on the plurality of contextual properties.
Claims17
A method for correcting/translating contextual language content comprising:
- identifying, by a processor, a first form of a first glyph having a plurality of contextual properties in the contextual language content;
- storing, by the processor, a plurality of buffer glyphs in a storage device, wherein the plurality of buffer glyphs surround the first glyph in the translated text;
- determining, by the processor, the plurality of contextual properties of the first glyph based on the plurality of buffer glyphs;
- and
- analyzing, by the processor, the first form of the first glyph with reference to one or more glyph form tables comprising a plurality of forms of the first glyph based, at least in part, on the plurality of contextual properties, wherein analyzing the first form of the first glyph comprises:
- determining, by the processor, a second form of the first glyph based, at least in part, on the one or more tables and the contextual properties;
- and
- indicating, by the processor, a corrected form of the first glyph.
A computer program product for correcting a symbol in a translated text comprising:
- a non-transitory, machine readable storage device having computer program instructions stored thereon for execution by a processor, the program instructions comprising:
- program instructions to identify a symbol having a plurality of contextual properties in a translation text;
- program instructions to store a plurality of buffer symbols in a storage device, wherein the plurality of buffer symbols surround the symbol in the translated text;
- program instructions to determine the contextual properties of the symbol based on the plurality of buffer symbols;
- program instructions to compare a first appearance of the symbol to one or more tables describing a plurality of appearances of the symbol based, at least in part, on the plurality of contextual properties;
- program instructions to determine a second appearance of the symbol based, at least in part, on the one or more tables and the contextual properties;
- and
- program instructions to replace the first appearance of the symbol with the second appearance of the symbol in the translated text.
A system for correcting a character in a translated text comprising:
- one or more processors;
- and
- a non-transitory, machine readable storage device having computer program instructions stored thereon for execution by a processor, the program instructions comprising:
- program instructions to identify a first style of a character having a plurality of contextual properties in a translation text;
- program instructions to store a plurality of environmental characters in a storage device, wherein the plurality of environmental characters surround the character in the translated text;
- program instructions to determine the contextual properties of the character based on the plurality of environmental characters;
- program instructions to compare the character to one or more data structures stored in the storage device and describing a plurality of styles of the character based, at least in part, on the plurality of contextual properties;
- program instructions to determine a second style of the character based, at least in part, on the one or more data structures and the contextual properties;
- and
- program instructions to replace the first style of the character with the second style of the character in the translated text.
Dependencies are inferred from the claim text.