Automated Scatter Plot Data Extraction via Template Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for extracting data from scatter plot images are cumbersome and error-prone, requiring manual input and lacking full automation, especially for large plots with diverse designs and artifacts, which complicates accurate data point and series identification.
Innovation Solution
A computer-implemented method that analyzes digital scatter plot images to identify pixel sets, generates templates for data points, and compares these templates with the image to automatically extract data points and their respective series, using techniques such as connected component analysis and template matching to enhance accuracy and robustness against errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual data entry is used for scatter plots, then data can be entered into the system, but the process is cumbersome and error-prone especially for large plots with dozens or hundreds of data points
Solution Approach 1:
The patent replaces manual mechanical data entry with automated image processing and pattern recognition systems. The system captures scatter plot images, processes them through algorithms to identify data points, and automatically extracts coordinate information, eliminating the need for manual keyboard input and significantly reducing both time and error rates.
Solution Approach 2:
The patent creates digital copies of scatter plot images and processes these copies to extract data. By working with image copies rather than requiring direct interaction with the original plot or manual transcription, the system enables automated data extraction that preserves accuracy while eliminating time-consuming manual entry processes.
2Extent of automation
If existing digitization programs are used, then some plot image processing is possible, but they do not fully-automatically extract data points or identify plot axes, maximum and minimum values, chart titles, series names, or axes labels
Solution Approach 1:
The patent segments the scatter plot image into distinct components including data points, axes, labels, titles, and legend elements. By dividing the image processing task into separate identification steps for each element type, the system achieves full automation while ensuring comprehensive extraction of all information elements without loss.
Solution Approach 2:
The patent performs preliminary identification and classification of various plot elements (axes, labels, titles, data points) before final data extraction. This preliminary action ensures that all elements are properly recognized and categorized, enabling complete automated extraction of data points, axis information, labels, titles, and series names without missing any critical information.
3Device complexity
If clustering algorithms are applied for identifying data points and their respective series in a single step, then the process is simplified, but overlapping data points may be erroneously identified as new types of data point symbols and new types of data series
Solution Approach 1:
The patent separates the identification process into distinct stages: first identifying data point locations, then determining series assignments. This segmentation prevents the conflation of overlapping points with new symbol types, as the system systematically processes each point through defined classification rules rather than relying on single-step clustering that may create erroneous categories.
Solution Approach 2:
The patent introduces an intermediary classification step between raw data point detection and final series assignment. This intermediary process uses reference templates and comparison algorithms to mediate the identification process, ensuring that overlapping points are correctly interpreted as belonging to existing series rather than being misidentified as new symbol types, thereby maintaining high reliability.
Data Source
AI summary
The invention relates to a computer-implemented method for automatically extracting data from a scatter plot. The method comprises receiving a digital image of a scatter plot; analyzing the received digital image for identifying a plurality of pixel sets, each pixel set being a group of adjacent pixels; analyzing the pixel sets in the received image or in a derivative of the received image for generating a plurality of templates; comparing the templates with pixels of a target image for identifying matching templates; identifying data points for the identified similar templates; assigning to each identified data point a data series; and returning the identified data points.