Systems and methods for handwriting recognition

A computing device for handwriting recognition, comprising:

• a memory storing instructions;
• a processor communicatively coupled with the memory and configured to:
• receive an input corresponding to a handwritten word having a plurality of characters;
• apply a first recognition model to the input handwritten word, the first recognition model configured to determine a first confidence level of a first portion of the input handwritten word and a second confidence level of a second portion of the input handwritten word, the first confidence level and the second confidence level indicating a likelihood of the first portion of the input handwritten word and the second portion of the input handwritten word including recognizable characters, wherein the first recognition model does not attempt to identify characters but identifies characteristics of the input handwritten word for potential characters;
• apply a second recognition model to the input handwritten word, wherein the second recognition model is different from the first recognition model, is selected based on the confidence levels determined by the first recognition model and is configured to determine characters of the input handwritten word based on the characteristics identified by the first recognition model and one or more words that include such characters;
• apply a third recognition model, wherein the third recognition model is configured to determine a list of candidate words based on the characteristics determined by the first recognition model and the words determined by the second recognition model;
• and
• output one or more candidate words for the input handwritten word from the list of candidate words.

A non-transitory computer-readable medium storing instructions for performing a method comprising:

• receiving an input corresponding to a handwritten word having a plurality of characters;
• applying a first recognition model to the input handwritten word, the first recognition model configured to determine a first confidence level of a first portion of the input handwritten word and a second confidence level of a second portion of the input handwritten word, the first confidence level and the second confidence level indicating a likelihood of the first portion of the input handwritten word and the second portion of the input handwritten word including recognizable characters, wherein the first recognition model does not attempt to identify characters but identifies characteristics of the input handwritten word for potential characters;
• applying a second recognition model to the input handwritten word, wherein the second recognition model is different from the first recognition model, is selected based on the confidence levels determined by the first recognition model, and is configured to determine characters of the input handwritten word based on the characteristics identified by the first recognition model and one or more words that include such characters;
• applying a third recognition model configured to determine a list of candidate words based on the characteristics determined by the first recognition model and the words determined by the second recognition model;
• and
• outputting one or more candidate words for the input handwritten word from the list of candidate words.