Learning Model for Handwriting Gradation Without Dedicated Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for adding gradation to electronically handwritten characters require dedicated devices for sensor information and pre-registered gradation patterns, limiting flexibility and creativity in expressing halftones.
Innovation Solution
An information processing system that uses a learning model to generate conversion image data, allowing for the addition of gradation to input images without dedicated devices or pre-defined patterns, by acquiring and binarizing partial images and learning from them to produce output images with added gradation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Shape
If sensor information (pressure value, voltage value) is used to add gradation to characters, then the handwriting-like appearance is improved, but dedicated devices are required which increases device complexity
Solution Approach 1:
The patent uses image reading to capture handwritten characters and creates digital copies that can be processed and enhanced with gradation effects. Instead of requiring dedicated sensing devices, the system reads the visual appearance of handwritten characters using a reading device, then applies learning models to generate gradation patterns that mimic the original handwriting style.
Solution Approach 2:
The patent replaces the mechanical/sensor-based system (pressure sensors, voltage sensors in dedicated pens) with an image processing system. By substituting physical sensing mechanisms with optical reading and computational learning, the system achieves similar handwriting-like effects without requiring specialized hardware.
2Productivity
If pre-registered gradation patterns are used to add gradation, then the processing speed is improved, but the versatility is reduced as only registered patterns can be applied
Solution Approach 1:
The patent implements a dynamic learning model that adapts to different handwriting styles and characteristics. Instead of using fixed, pre-registered gradation patterns, the system learns from input images and generates appropriate gradation patterns in real-time, allowing it to handle diverse handwriting styles while maintaining efficient processing through learned patterns.
Solution Approach 2:
The patent changes the parameters of gradation patterns based on the learned characteristics of the input handwriting. The learning model adjusts gradation parameters (such as tone distribution, pattern density, and intensity) according to the specific features of each handwritten character, enabling versatile application across different writing styles without requiring pre-registration of multiple patterns.
3Measurement precision
If dedicated devices with sensors are used to acquire handwriting information, then the measurement precision of pressure and speed is improved, but the ease of operation is reduced due to specialized equipment requirements
Solution Approach 1:
The patent enables the system to extract necessary handwriting characteristics (pressure, speed, stroke patterns) from the visual information in the read image itself. The learning model automatically analyzes the input image to infer handwriting dynamics without requiring external sensor data, making the system self-sufficient and easier to operate with ordinary writing tools.
Solution Approach 2:
The patent creates a universal system that can process handwriting from any ordinary writing tool (pen, pencil, brush) without requiring specialized sensor-equipped devices. The image reading and learning approach works with diverse writing instruments, making the system multi-functional and accessible to general users with common writing materials.
Data Source
AI summary
An information processing system acquires, using a reading device, a read image from an original on which a handwritten character is written; acquires, based on the read image, a partial image that is a partial region of the read image and a binarized image that expresses the partial image by two tones; performs learning of a learning model based on learning data that uses the partial image as a correct answer image and the binarized image as an input image; acquires print data including a font character; generates conversion image data including a gradation character obtained by inputting the font character to the learning model; and causes an image forming device to form an image based on the generated conversion image data.