Inferring Stroke Information from Hand-Drawn Text Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Intelligent character recognition (ICR) performance suffers when applied to archived images of hand-drawn text characters, as timing information is not available, leading to reduced accuracy and efficiency in recognizing these characters.
Innovation Solution
A method and system that extract character segments from an image, determine character bounding boxes, and analyze texture properties, brush widths, and intensities to provide ordering and direction information to an ICR engine, enhancing the recognition process by utilizing these properties to improve matching and reduce processing time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If ICR is performed on archived images without timing information, then the technology can recognize text characters in stored images, but the recognition accuracy and performance suffer
Solution Approach 1:
The patent performs preliminary analysis of the archived image to extract timing information from texture properties, brush widths, and intensities before submitting to the ICR engine. This preliminary extraction of directional and timing data compensates for the absence of original timing information, thereby improving recognition accuracy without requiring the original timing data to be present
Solution Approach 2:
The patent introduces an intermediary processing step that analyzes texture properties, brush widths, and intensities to infer timing information. This intermediary analysis acts as a mediator between the archived image data and the ICR engine, translating static image properties into temporal characteristics that improve recognition performance
2Productivity
If ICR is performed without timing information, then archived images can be processed, but processing efficiency is reduced
Solution Approach 1:
The system performs preliminary extraction of directional and timing information from texture properties and brush characteristics before the main ICR processing. This advance preparation reduces the computational burden during actual recognition, thereby improving processing efficiency and reducing output time
3Measurement precision
If traditional ICR is used on hand-drawn characters, then the process is simple, but matching accuracy is poor
Solution Approach 1:
The patent segments the hand-drawn character into character segments and determines character bounding boxes, analyzing texture properties and brush widths for each segment. This segmentation allows for more precise local analysis of drawing characteristics, improving matching accuracy while managing complexity through structured processing
Solution Approach 2:
The patent transforms static image parameters (texture properties, brush widths, intensities) into dynamic timing and directional information. By changing the representation of these parameters from spatial to temporal-domain characteristics, the system improves matching accuracy for hand-drawn characters
Data Source
AI summary
A method for character recognition. The method includes: obtaining a plurality of character segments extracted from an image; determining a first character bounding box including a first set of the plurality of character segments and a second character bounding box including a second set of the plurality of character segments; determining an ordering for the first set based on a plurality of texture properties for the first set; determining a plurality of directions of the first set based on a plurality of brush widths and a plurality of intensities for the first set; and executing character recognition for the first character bounding box by sending the first set, the plurality of directions for the first set, and the ordering for the first set to an intelligent character recognition (ICR) engine.