Touch Panel Shape Identification Using Trace Graph Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for identifying complex shapes drawn on touch panels face challenges in accurately extracting line factors and estimating the number of vertices, leading to imprecise comparisons with ideal reference graphics, especially for spiral-shaped graphics where angle histograms fail to differentiate between shapes.
Innovation Solution
A graphics identification program that detects traces on a touch panel, calculates line segment lengths and angles, and compares them to reference graphics in an orthogonal coordinate system, using a step graph representation to determine similarity based on cumulative length and area differences, allowing for accurate identification of complex shapes regardless of vertex number accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the number of vertices is estimated using angle histogram, then the identification process is simple, but the accuracy of identifying complex shapes deteriorates
Solution Approach 1:
The patent transforms the graphic identification from vertex-based counting to a continuous curve analysis approach. By representing graphics as sequences of coordinate points with calculated angles and distances, the system moves from discrete vertex estimation to continuous shape characterization, enabling accurate identification of complex shapes without relying on vertex number estimation.
Solution Approach 2:
The patent changes the identification parameters from vertex count to angle sequences and distance sequences. By calculating the angle between adjacent line segments and the distance between vertices, the system creates a detailed parameter set that uniquely characterizes each graphic shape, allowing precise differentiation even for complex shapes with many vertices.
2Loss of information
If line factors are extracted from coordinate points, then vertex estimation can be performed, but extraction accuracy deteriorates for complex shapes
Solution Approach 1:
The patent performs preliminary calculations of angle and distance parameters for each coordinate point before conducting the identification comparison. By pre-calculating the angle between adjacent line segments and the distance between vertices for all points in the sequence, the system prepares accurate reference data that captures the complete shape characteristics, eliminating the need for post-hoc vertex estimation.
3Productivity
If angle histogram is used for vertex estimation, then processing is simplified, but the ability to differentiate spiral shapes is lost
Solution Approach 1:
The patent segments the graphic into individual coordinate points in chronological order, calculating the angle and distance for each segment between adjacent points. This segmentation approach preserves the sequential structure and directional information of the drawing, allowing the system to differentiate spiral shapes based on the continuous variation of angle and distance parameters along the path.
Data Source
AI summary
An information processing device executing a graphics identification program detects a trace drawn on an input surface as coordinate points in chronological order. Coordinate points for representing the trace are selected from a group of coordinate points arranged in chronological order. Lengths of line segments obtained by connecting the selected coordinate points in chronological order are calculated. Also, for each line segment, an absolute angle and a relative angle are calculated. A trace graph in an orthogonal coordinate system, whose first axis indicates a line segment distance and whose second axis indicates an angle of a line segment whose startpoint is one of the selected coordinate points, is compared with a reference graphic graph representing a reference graphic in the orthogonal coordinate system, thereby identifying a reference graphic represented by a reference graphic graph which most closely resembles the trace graph as a shape of the trace.