Node-Scored OCR for Low-Resolution Vehicle Text Recognition

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Manually identifying vehicle identification numbers is difficult due to their large volume and varied locations, making automated recognition more efficient for various uses.

Innovation Solution

A system utilizing a machine learning model trained with augmented digital images to detect vertically and horizontally oriented text on objects, employing a pattern agnostic optical character recognition (OCR) system to accurately identify characters by grouping and scoring nodes within the same location, and transmitting recognized characters for display.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If automated recognition systems are implemented to identify vehicle identification numbers, then productivity and efficiency are improved, but device complexity increases due to the need for sophisticated OCR and machine learning models

Engineering Contradiction:
Improveidentification efficiencyVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system segments the text recognition task into distinct processing stages: image capture, preprocessing, node detection, character grouping, and recognition. This segmentation allows each component to be optimized independently while working together to achieve accurate vehicle identification number recognition, resolving the contradiction between automation benefits and system complexity

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The OCR system is designed with universal capabilities to handle multiple text orientations (vertical and horizontal), various font types, and different image qualities. This multi-functionality enables a single system to address diverse vehicle identification scenarios without requiring separate specialized systems for each case, improving productivity while managing complexity through consolidation

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Use of energy by moving object

If the system processes images at lower resolutions to reduce data processing requirements, then use of energy and processing time are improved, but measurement precision deteriorates due to reduced character recognition accuracy

Engineering Contradiction:
Improveprocessing energyVSAvoidcharacter recognition accuracy
Core Design Contradiction:
Use of energy by moving objectVSMeasurement precision

Solution Approach 1:

The system performs preliminary preprocessing operations on images before full OCR processing, including normalization, contrast enhancement, and noise reduction. These preliminary actions prepare lower-resolution images for accurate processing by enhancing critical features, allowing the system to maintain recognition accuracy while operating at reduced resolution to conserve energy

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts processing parameters based on image characteristics, applying different levels of preprocessing and node detection sensitivity. For lower-resolution images, the system modifies parameters such as node size thresholds and grouping criteria to optimize recognition accuracy, enabling energy-efficient processing without sacrificing measurement precision

Inventive Principle:
Principle #35Parameter changes

3Reliability

If the system maintains high accuracy in character differentiation, then reliability is improved, but device complexity increases due to the need for sophisticated node grouping and scoring mechanisms

Engineering Contradiction:
Improvecharacter identification accuracyVSAvoidrecognition system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system applies local quality analysis by examining specific regions and characteristics of characters individually. Each character is evaluated based on its local node patterns, spatial relationships, and contextual features. This localized approach enables high accuracy in character differentiation without requiring complex global analysis of entire text strings, maintaining reliability while managing system complexity

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS12374137B2Systems and methods for recognizing text of interest
Publication Date: 2025.07.29 WALMART APOLLO LLC
  • US12374137B2 patent drawing
  • US12374137B2 patent drawing
  • US12374137B2 patent drawing

AI summary

In some embodiments, apparatuses and methods are provided herein useful to determine text on an object. In some embodiments, there is provided a system to determine text of interest on an object of interest including a control circuit configured to execute a machine learning model trained to identify the text of interest, group into a cluster each character in the text of interest located substantially in the same location in the text of interest, determine a score value of each particular character in the cluster, identify the particular character that has a determined score value corresponding to at least a threshold score value relative to all characters in the cluster, assign the particular character having the determined score value corresponding to at least the threshold score value as a recognized character in the cluster, and output data comprising each recognized character associated with each cluster in the text of interest.