# US 2024/0203147 A1 · Systems and methods for digitized document image text contouring > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2024/0203147 A1 - Document ID: US20240203147A1 - Status: Published application - Kind: A1 - Application number: 18/530,481 - Filed: 2023-12-06 - Published: 2024-06-20 - Earliest claimed: 2020-01-21 - Assignees: Capital One Services, LLC - Applicants: Capital One Services, LLC - Inventors: Douglas SLATTERY - Web page: https://patentgrep.com/patent/US20240203147A1 ## CPC - G06V 30/412: Layout analysis of documents structured with printed lines or input boxes, e.g. business forms or tables - G06V 10/44: Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components - G06V 20/62: Text, e.g. of license plates, overlay texts or captions on TV images - G06V 30/153: {using recognition of characters or words} - G06V 30/414: Extracting the geometrical structure, e.g. layout tree; Block segmentation, e.g. bounding boxes for graphics or text - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract Systems and methods for digitized document image text contouring are provided. One or more memories may be coupled to one or more processors, the one or more memories including instructions operable to be executed by the one or more processors. The one or more processors may be configured to receive a digitized document image. The one or more processors may be configured to preprocess the digitized document image to generate a plurality of contours. The one or more processors may be configured to adjust a plurality of bounding boxes of the plurality of contours; analyze the adjusted plurality of bounding boxes; create one or more clips based on the analysis; dynamically threshold the one or more clips; perform optical character recognition of the one or more clips; and receive output responsive to the optical character recognition. ## Claims ### Claim at position 1 1-41. (canceled) ### Claim 42 A method for digitized document image text contouring, comprising: receiving, by a processor, a digitized document image; performing, by the processor, skew adjustment or morphage of the digitized document image for text contouring; applying, by the processor, a first pass over the digitized document image to generate a first contour; applying, by the processor, a second pass over the digitized document image to generate a second contour; adjusting, by the processor, a bounding box of the first contour for data spillage outside of the bounding box of the first contour; adjusting, by the processor, a bounding box of the second contour for data spillage outside of the bounding box of the second contour; combining, by the processor, the bounding box of the first contour and the bounding box of the second contour; producing a clip, by the processor, based on combining the bounding box of the first contour and the bounding box of the second contour; dynamically thresholding, by the processor, the clip; transmitting, by the processor, the clip for optical character recognition; and receiving, by the processor, an output responsive to the optical character recognition. ### Claim 43 (depends on 42) The method of claim 42, where the action of performing, by the processor, skew adjustment or morphage of the digitized document image for text contouring, comprises: utilizing, by the processor, edge detection to the digitized document image to determine a linearity of the digitized document image based on vertical and horizontal edges. ### Claim 44 (depends on 43) The method of claim 43, wherein the action of utilizing, by the processor, edge detection to the digitized document image to determine a linearity of the digitized document image based on vertical and horizontal edges, comprises: identifying, via canny edge detection, a vertical edge of the digitized document image and a horizontal edge of the digitized document image; identifying, via canny edge detection one or more boundaries of the digitized document image; calculating a first degree of rotation between the vertical edge of the digitized document image and at least one of the one or more boundaries of the digitized document image; and calculating a second degree of rotation between the horizontal edge of the digitized document image and at least one of the one or more boundaries of the digitized document image. ### Claim 45 (depends on 42) The method of claim 42, further comprising: applying, by the processor, an Otsu threshold filter over the digitized document image to adjust one or more of brightness and contrast to improve readability of the digitized document image. ### Claim 46 (depends on 42) The method of claim 42, further comprising: applying, by the processor, a despeckle filter over the digitized document image to remove speckles from the digitized document image such that one or more pixels are whited out. ### Claim 47 (depends on 42) The method of claim 42, wherein the first contour and the second contour comprise a line or a shape, or a portion of a line or shape, including a regular or irregular form. ### Claim 48 (depends on 42) The method of claim 42, wherein the first pass is configured for contouring one or more preprinted boxes in the digitized document image. ### Claim 49 (depends on 42) The method of claim 42, wherein the first pass includes canny edge detection to identify one or more shapes in the digitized document image. ### Claim 50 (depends on 42) The method of claim 42, wherein the second pass includes a dilation filter configured to prepare the digitized document image to capture free formed printed text and/or inputted text. ### Claim 51 (depends on 42) The method of claim 42, wherein the second pass is configured to validate or confirm text as-is. ### Claim 52 A system, comprising: a processor; and a memory storing a digitized document image, wherein the processor is configured to: receive the digitized document image; perform skew adjustment or morphage of the digitized document image for text contouring; apply a first pass over the digitized document image to generate a first contour; apply a second pass over the digitized document image to generate a second contour; adjust a bounding box of the first contour for data spillage outside of the bounding box of the first contour; adjust a bounding box of the second contour for data spillage outside of the bounding box of the second contour; combine the bounding box of the first contour and the bounding box of the second contour; produce a clip based on combining the bounding box of the first contour and the bounding box of the second contour; dynamically threshold the clip; transmit the clip for optical character recognition; and receive an output responsive to the optical character recognition. ### Claim 53 The system of 52, wherein the second pass includes canny edge detection to identify one or more shapes in the digitized document. ### Claim 54 The system of 53, wherein the one or more shapes are smudged together to create one or more blobs of pixels. ### Claim 55 The system of 53, wherein the canny edge detection yields one or more coordinate pairs to derive one or more geometric point arrays for creating the bounding box of the second contour. ### Claim 56 The system of 52, wherein the processor is further configured to identify a minimum and maximum pixel area range that the first and second contours desirably fall within. ### Claim 57 The system of 52, wherein the action of dynamically threshold the clip comprises altercating visibility of one or more pixels in the clip. ### Claim 58 The system of 52, wherein the action of dynamically threshold the clip comprises applying a formula to determine a optimal threshold for achieving maximum optical character recognition (OCR) accuracy, in part, based on frequency of the most dark pixels in the clip. ### Claim 59 The system of 52, wherein the output comprises textual output as a result of OCR. ### Claim 61 The system of 52, wherein the action of combine the bounding box of the first contour and the bounding box of the second contour, comprises merging the bounding box of the first contour and the bounding box of the second contour to eliminate encapsulated boxes and overlaps. ### Claim 62 A computer readable non-transitory medium comprising computer-executable instructions that, when executed by a digitized document image text contouring system, cause the digitized document image text contouring system to perform operations of: receiving a digitized document image; performing skew adjustment or morphage of the digitized document image for text contouring; applying a first pass over the digitized document image to generate a first contour; applying a second pass over the digitized document image to generate a second contour; adjusting a bounding box of the first contour for data spillage outside of the bounding box of the first contour; adjusting a bounding box of the second contour for data spillage outside of the bounding box of the second contour; combining the bounding box of the first contour and the bounding box of the second contour; producing a clip based on combining the bounding box of the first contour and the bounding box of the second contour; dynamically thresholding the clip; transmitting the clip for optical character recognition; and receiving an output responsive to the optical character recognition. ### Claim 560 The system of 52, wherein the process is configured to perform one or more conditional checks to validate one or more characters by transmitting the clip through one or more additional OCR engines.