Image preprocessing for optical character recognition

A machine-implemented method for processing images, the method comprising:

• receiving a captured image that contains:
• (a) a region of interest (ROI) including a plurality of characters to be autonomously recognized as text and included in an output of an optical character recognition (OCR) system, and (b) non-ROI content to be excluded from the OCR system output;
• autonomously processing the captured image to detect and locate the ROI in the captured image, and to determine a boundary of the ROI, wherein the processing includes:
• performing a transformation of the captured image to a first feature descriptor representation (FDR) of the captured image;
• and
• performing comparison between the first FDR and at least one ROI template that includes at least a second FDR of a representative ROI image;
• wherein the processing produces an output to be provided to an OCR engine to perform autonomous OCR processing of the ROI while ignoring the non-ROI content based on the determined boundary of the ROI, wherein the processing includes performing a plurality of comparisons between multiple FDRs of respective multiple differently-scaled variants of the captured image and the at least one ROI template, to produce a corresponding plurality of comparison results, wherein the plurality of comparison results are stored in a multi-dimensional data structure wherein comparison results represented along a first dimension correspond to comparisons using the multiple differently-scaled variants of the captured image, and wherein comparison results represented along a second dimension correspond to comparisons using multiple rotated variants of an angular alignment between each of the multiple FDRs of the multiple differently-scaled variants of the captured image and the second FDR.

At least one non-transitory machine-readable medium containing instructions for processing images, wherein the instructions when executed on a computing system, cause the computing system to:

• receive a captured image that contains:
• (a) a region of interest (ROI) including a plurality of characters to be autonomously recognized as text and included in an output of an optical character recognition (OCR) system, and (b) non-ROI content to be excluded from the OCR system output;
• perform autonomously processing of the captured image to detect and locate the ROI in the captured image, and to determine a boundary of the ROI, wherein the processing includes:
• performing a transformation of the captured image to a first feature descriptor representation (FDR) of the captured image;
• and
• performing comparison between the first FDR and at least one ROI template that includes at least a second FDR of a representative ROI image to produce a corresponding plurality of comparison results stored in a data structure, at least some of the stored comparison results corresponding to comparisons using multiple differently-scaled variants of the captured image or the at least one ROI template, and at least some of the stored comparison results corresponding to comparisons using multiple rotated variants of an angular alignment between the captured image and the at least one ROI image;
• producing an output to be provided to an OCR engine to perform autonomous OCR processing of the ROI while ignoring the non-ROI content based on the determined boundary of the ROI.