Systems and methods for enhancing trainable optical character recognition (OCR) performance

A machine vision system for enhancing trainable optical character recognition (OCR) performance, the system comprising:

• a machine vision camera configured to capture an image including a barcode encoding a payload and a character string;
• and
• a user computing device executing an application, wherein the user computing device is communicatively coupled to the machine vision camera and is configured to:
• receive the image captured by the machine vision camera,
• identify, in the image, the barcode and the character string,
• decode the barcode to determine the payload,
• apply an optical character recognition (OCR) algorithm to the image, wherein the OCR algorithm interprets the character string and identifies an unrecognized character within the character string,
• compare the payload to the character string to validate the unrecognized character as corresponding to a known character included within the payload, and
• responsive to validating the unrecognized character, add the unrecognized character to a font library referenced by the OCR algorithm as an entry for the known character;
• wherein the user computing device is further configured to:
• determine, by the OCR algorithm, whether or not the unrecognized character satisfies an acceptable confidence threshold when compared to the known character included within the payload,
• responsive to determining that the unrecognized character does not satisfy the acceptable confidence threshold, compare, by the OCR algorithm, each entry in the font library to the unrecognized character,
• determine, by the OCR algorithm, whether or not the unrecognized character satisfies a respective font library entry threshold, wherein the respective font library entry threshold is based on a respective composite similarity score between the unrecognized character and each entry of the font library corresponding to the respective known character, and
• responsive to determining that the unrecognized character satisfies a first respective font library entry threshold, validate the unrecognized character as the first respective known character included within the payload.

A tangible machine-readable medium comprising instructions that, when executed, cause a machine to at least:

• receive, from a machine vision camera, an image including a barcode encoding a payload and a character string;
• identify, in the image, the barcode and the character string;
• decode the barcode to determine the payload;
• apply an optical character recognition (OCR) algorithm to the image, wherein the OCR algorithm interprets the character string and identifies an unrecognized character within the character string;
• compare the payload to the character string to validate the unrecognized character as corresponding to a known character included within the payload;
• and
• responsive to validating the unrecognized character, add the unrecognized character to a font library referenced by the OCR algorithm as an entry for the known character;
• wherein the instructions, when executed, further cause the machine to at least:
• determine, by the OCR algorithm. whether or not the unrecognized character satisfies an acceptable confidence threshold when compared to the known character included within the payload;
• responsive to determining that the unrecognized character does not satisfy the acceptable confidence threshold, compare, by the OCR algorithm, each entry in the font library to the unrecognized character;
• determine, by the OCR algorithm, whether or not the unrecognized character satisfies a respective font library entry threshold, wherein the respective font library entry threshold is based on a respective composite similarity score between the unrecognized character and each entry of the font library corresponding to the respective known character;
• and
• responsive to determining that the unrecognized character satisfies a first respective font library entry threshold, validate the unrecognized character as the first respective known character included within the payload.