Traffic Sign Character Recognition via Morphological Thinning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing traffic sign recognition systems rely on computationally complex machine learning techniques, which are sensitive to illumination and scale, and often fail to accurately recognize incomplete digits, leading to incorrect recognition.

Innovation Solution

A system that uses morphological preprocessing and a template-based approach to recognize alphanumeric characters from traffic signs, independent of machine learning, by converting images to binary, applying thinning, and comparing pixel counts against pre-determined thresholds, invariant to lighting conditions and scale.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If machine learning techniques are used for traffic sign recognition, then recognition capability is improved, but computational complexity increases and processing time increases

Engineering Contradiction:
Improverecognition capabilityVSAvoidcomputational complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the traffic sign recognition process into distinct stages: traffic sign detection, character extraction, and character recognition. Each stage uses specialized algorithms optimized for its specific task, avoiding the need for a single complex machine learning model to handle all aspects of recognition.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent replaces machine learning-based recognition with a template matching approach that uses geometric and topological features. This substitution reduces computational complexity by using deterministic geometric transformations and pixel count comparisons instead of probabilistic machine learning models.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If machine learning techniques are used for traffic sign recognition, then recognition capability is improved, but processing time increases

Engineering Contradiction:
Improverecognition capabilityVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent performs preliminary actions by detecting traffic signs and extracting characters before recognition. By pre-processing the image to isolate characters and convert to binary format, the system reduces the computational burden during the recognition phase, enabling faster processing compared to end-to-end machine learning approaches.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent substitutes time-consuming machine learning inference with faster template matching and geometric feature comparison. The use of pixel count calculations and template correlation provides rapid recognition results without the iterative computation required by neural networks.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Device complexity

If traditional recognition methods are used, then computational complexity is reduced, but recognition accuracy deteriorates due to sensitivity to illumination and scale

Engineering Contradiction:
Improvecomputational complexityVSAvoidrecognition accuracy
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent changes parameters by converting images to binary format and applying geometric normalization. This transformation makes the recognition process invariant to illumination variations and scale changes, allowing accurate recognition while maintaining low computational complexity.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent uses asymmetric template matching where the template represents the ideal character form and the extracted character is compared against it. This asymmetric comparison approach, combined with pixel count normalization, enables accurate recognition regardless of scale or lighting conditions while keeping the algorithm simple.

Inventive Principle:
Principle #4Asymmetry

4Adaptability or versatility

If scale-variant recognition is used, then adaptability to different image sizes is improved, but measurement precision deteriorates due to incomplete digit extraction

Engineering Contradiction:
Improvescale invarianceVSAvoiddigit extraction accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent applies parameter changes by converting extracted characters to binary format and normalizing their size through geometric transformations. This standardization ensures that characters from different scales are represented uniformly, enabling precise pixel count comparison and accurate recognition while maintaining scale adaptability.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3051461B1A computer implemented system and method for extracting and recognizing alphanumeric characters from traffic signs
Publication Date: 2020.04.15 TATA CONSULTANCY SERVICES LTD
  • EP3051461B1 patent drawingFigure 1
  • EP3051461B1 patent drawingFigure 2(a)~2(c)
  • EP3051461B1 patent drawingFigure 3(a)~3(b)

AI summary

A computer implemented system and method for extracting and recognizing alphanumeric characters from traffic signs is envisaged. A camera present in the system is adapted to be mounted on the vehicle for capturing images of traffic signs. The system then gives a region of interest (ROI) within the frame of the captured image, from where the alphanumeric characters are extracted and enhanced for recognition. For recognizing the enhanced alphanumeric characters, the image is converted to binary image and morphological thinning technique is applied on it. The space then obtained between two alphanumeric characters thus helps in extracting the characters and recognizing them independently from the captured image. For each extracted alphanumeric character is re-sized to a specified size and is recognized with the help of white pixels count. Once the alphanumeric characters are recognized they are displayed on a display.