Text recognition in a video stream using a combination of recognition results with per-character weighting

A system comprising:

• at least one hardware processor;
• and
• one or more software modules that are configured to, when executed by the at least one hardware processor,
• receive a plurality of image frames from a video of a document,
• for each of the plurality of image frames,
• perform text recognition on the image frame to obtain a text-recognition result for the image frame, wherein the text-recognition result comprises a plurality of character-recognition results, wherein each character-recognition result represents a character in a text field in the document, and
• calculate a frame weight for the image frame, and
• combine the text-recognition results for at least a subset of the plurality of image frames by
• aligning the character-recognition results across the text-recognition results for the at least a subset of the plurality of image frames into an alignment with a plurality of positions, wherein each of the plurality of positions represents a position of a character in the text field in the document,
• calculating a character weight for each character-recognition result in the text-recognition results for the at least a subset of the plurality of image frames, and,
• at each of the plurality of positions in the alignment, accumulating the character-recognition results that are at that position, across the text-recognition results for the at least a subset of the plurality of image frames, based on the character weights and the frame weights, to produce an accumulated text-recognition result that represents the text field in the document.

A non-transitory computer-readable medium having instructions stored therein, wherein the instructions, when executed by a processor, cause the processor to:

• receive a plurality of image frames from a video of a document;
• for each of the plurality of image frames,
• perform text recognition on the image frame to obtain a text-recognition result for the image frame, wherein the text-recognition result comprises a plurality of character-recognition results, wherein each character-recognition result represents a character in a text field in the document, and
• calculate a frame weight for the image frame;
• and
• combine the text-recognition results for at least a subset of the plurality of image frames by
• aligning the character-recognition results across the text-recognition results for the at least a subset of the plurality of image frames into an alignment with a plurality of positions, wherein each of the plurality of positions represents a position of a character in the text field in the document,
• calculating a character weight for each character-recognition result in the text-recognition results for the at least a subset of the plurality of image frames, and,
• at each of the plurality of positions in the alignment, accumulating the character-recognition results that are at that position, across the text-recognition results for the at least a subset of the plurality of image frames, based on the character weights and the frame weights, to produce an accumulated text-recognition result that represents the text field in the document.