US 11,928,875 B2
Layout-aware, scalable recognition system
Yan Wang, Mercer Island, WA (US); Ye Wu, Bothell, WA (US); and Arun Sacheti, Sammamish, WA (US)
Assigned to MICROSOFT TECHNOLOGY LICENSING, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Mar. 8, 2019, as Appl. No. 16/297,388.
Prior Publication US 2020/0285878 A1, Sep. 10, 2020
Int. Cl. G06F 16/387 (2019.01); G06F 16/31 (2019.01); G06F 16/35 (2019.01); G06F 18/2411 (2023.01); G06N 20/10 (2019.01); G06V 10/70 (2022.01); G06V 10/75 (2022.01); G06V 10/80 (2022.01); G06V 20/62 (2022.01); G06V 30/144 (2022.01); G06V 30/148 (2022.01); G06V 30/18 (2022.01); G06V 30/19 (2022.01); G06V 30/413 (2022.01); G06V 30/414 (2022.01); G06V 30/10 (2022.01)
CPC G06V 30/18171 (2022.01) [G06F 16/313 (2019.01); G06F 16/35 (2019.01); G06F 16/387 (2019.01); G06F 18/2411 (2023.01); G06N 20/10 (2019.01); G06V 10/70 (2022.01); G06V 10/75 (2022.01); G06V 10/806 (2022.01); G06V 20/62 (2022.01); G06V 30/144 (2022.01); G06V 30/153 (2022.01); G06V 30/158 (2022.01); G06V 30/1916 (2022.01); G06V 30/19173 (2022.01); G06V 30/413 (2022.01); G06V 30/414 (2022.01); G06V 30/10 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A method for visual recognition, comprising:
receiving an output of an optical character recognition process comprising a recognized block, the recognized block comprising a set of positional properties and recognized text;
creating a block weight from the set of positional properties for the recognized block;
creating a set of block tri-character grams for the recognized block based on the recognized text for the recognized block;
creating a block tri-character gram feature vector based on the set of block tri-character grams;
creating a block feature vector for the recognized block from the block weight and the block tri-character gram feature vector for the recognized block;
creating a feature vector based on the block feature vector;
retrieving a set of results from an index, wherein the index includes entries for documents, and further wherein the set of results is retrieved based on the feature vector; and
submitting the feature vector to a trained machine classifier,
wherein retrieving the set of results from the index based on the feature vector occurs responsive to the trained machine classifier classifying the feature vector into a designated category.