# US 12,762,877 B2 · Computer-implemented segmented numeral character recognition and reader > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,762,877 B2 - Document ID: US12762877B2 - Status: Granted - Kind: B2 - Application number: 18/669,750 - Filed: 2024-05-21 - Published: 2024-09-12 - Granted: 2026-10-13 - Earliest claimed: 2021-12-09 - Assignees: Redimd, LLC - Applicants: Redimd, LLC - Inventors: Walter Holmsten; Malcolm Edward Storey; Aakash L. Gupta - Web page: https://patentgrep.com/patent/US12762877B2 ## CPC - G06V 30/19147: {Obtaining sets of training patterns; Bootstrap methods, e.g. bagging or boosting} - G06V 10/945: {User interactive design; Environments; Toolboxes} - G06V 20/62: Text, e.g. of license plates, overlay texts or captions on TV images - G06V 30/147: {Determination of region of interest} - G06V 30/19007: {Matching; Proximity measures} - G06V 30/19107: {Clustering techniques} - G06V 30/1916: {Validation; Performance evaluation} - G16H 30/20: for handling medical images, e.g. DICOM, HL7 or PACS - G16H 40/67: for remote operation - G06V 2201/02: Recognising information on displays, dials, clocks - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING - G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA ## Abstract Computer-implemented methods, systems and devices having segmented numeral character recognition. In an embodiment, users may take digital pictures of a seven-segment display on a sensor device. For example, a user at a remote location may use a digital camera to capture a digital image of a seven-segment display on a sensor device. Captured images of a seven-segment display may then be sent or uploaded over a network to a remote health management system. The health care management system includes a reader that processes the received images to determine sensor readings representative of the values output on the seven-segment displays of the remote sensor devices. Machine learning and OCR are used to identify numeric characters in images associated with seven-segment displays. In this way, a remote heath management system can obtain sensor readings from remote locations when users only have access to sensor devices with seven-segment displays. ## Claims ### Claim 1 A computer-implemented system for numeric character recognition of numerals shown in a segmented display on a sensor device over a data network, comprising: means for capturing a digital image of a segmented display of numeric values shown on a sensor device; means for sending the digital image to a remote health management system having a reader; means for storing the digital image in a computer-readably memory; and means for processing the digital image at the reader to determine a sensor reading representative of the numeric values in the segmented display of the sensor device, wherein the processing means includes: means for analyzing the digital image using optical character recognition to determine a first array of data; means for analyzing the digital image using a trained machine learning (ML) model to determine a second array of data; and means for evaluating the first and second arrays of data to obtain the sensor reading representative of the numeric values in the segmented display of the sensor device. ### Claim 2 (depends on 1) The computer-implemented system of claim 1, wherein the segmented display comprises a seven-segment display where seven segments are displayed on or off to represent particular numeric values between 0 and 9. ### Claim 3 (depends on 1) The computer-implemented system of claim 1, wherein the processing means includes means for identifying a region of interest and clustering digits. ### Claim 4 (depends on 3) The computer-implemented system of claim 3, wherein the identifying a region of interest means includes means for object detection. ### Claim 5 (depends on 1) The computer-implemented system of claim 1, further comprising means for training a model having parameters, and wherein the processing means includes means for minimizing a loss function and hyperparameter tuning of parameters in the model. ### Claim 6 (depends on 1) The computer-implemented system of claim 1, wherein the evaluating means includes: means for comparing the first and second arrays of data to determine whether data in the first and second arrays is identical or empty; means for outputting an output array of data having data from the first or second array of data when identical and from the first or second array that is not empty; and means for arbitrating the first and second arrays of data according to test criteria to obtain an output array of data when the comparing means determines the first and second arrays of data are not identical and empty. ### Claim 7 (depends on 6) The computer-implemented system of claim 6, further comprising means for temporarily saving the obtained output array of data in computer-readable memory along with an identifier associated with the digital image. ### Claim 8 (depends on 1) The computer-implemented system of claim 1, further comprising: means for displaying the sensor reading information as numeric values along with the associated digital image for confirmation; and means for enabling a user to edit or select the numeric values and submit an input indicative of user confirmation. ### Claim 9 (depends on 8) The computer-implemented system of claim 8, further comprising: means for comparing the user confirmation input and the temporarily saved output array of data; means for checking whether the comparison is passed or not passed; means for flagging the sensor reading information of the output array of data for administrative review when the comparison is not passed; and means for saving the sensor reading information of the output array of data in a record of a database. ### Claim 10 (depends on 1) The computer-implemented system of claim 1, further comprising means for training the ML engine with a training dataset of images to obtain the trained ML model. ### Claim 11 (depends on 10) The computer-implemented system of claim 10, wherein the training dataset includes base images of different sensor reading devices having segmented numeral displays, and the training means includes: means for augmenting base images in the training dataset to generate a synthetic training dataset that includes base augmentations representative of one or more of the following augmentations: Horizontal Flip, Vertical Flip, Shift-Scale-Rotate, Random-Brightness-Contrast, Random Sun Flare, Random Fog, or Blur of each base image in the training dataset. ### Claim 12 (depends on 10) The computer-implemented system of claim 10, wherein the training means includes: means for applying images in the training dataset to a ML engine; means for evaluating candidate models using objection detection, means for minimizing a loss function and hyper-parameters to obtain the trained ML model. ### Claim 13 (depends on 1) The computer-implemented system of claim 1, wherein the analyzing means for the digital image using the trained ML model includes means for passing to the digital image with an API call to a Detectron2Go library to determine the second array of data.