Automatic Line Erase Tool for Patent Drawing Annotation Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for processing patent drawings are inefficient due to the presence of extraneous information like lead lines and part numbers, which hinder semantic processing and manual derivation of information, making it difficult to focus on desired information and convert 2D drawings to 3D models.
Innovation Solution
A method involving linear regression analysis to detect and remove lead lines by tracing their path and updating points, allowing for the removal of extraneous information and enhancing the semantic value of drawings for machine learning and 2D to 3D conversion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If patent drawings are published in digital image format without semantic processing, then the drawings can be displayed and viewed, but the information cannot be readily harvested or processed automatically
Solution Approach 1:
The patent extracts lead lines and part numbers as separate identifiable elements from the drawing image. By detecting lead lines through linear regression analysis and tracing their paths from part numbers to drawing features, the system separates annotation information from the actual drawing content, enabling automatic information harvestability without requiring full semantic processing of the entire image.
Solution Approach 2:
The patent segments the drawing information into distinct components: lead lines, part numbers, and drawing features. Each component is processed independently through specific detection algorithms, allowing the system to handle different types of information separately and improve overall automatic processing capability while maintaining information integrity.
2Loss of information
If lead lines and part numbers are retained in patent drawings, then the drawings provide complete annotation information, but the drawings become crowded and difficult to focus on desired information
Solution Approach 1:
The system extracts lead lines and part numbers as separate detectable elements, allowing them to be selectively removed or highlighted based on user needs. This extraction enables the creation of simplified drawing versions that maintain essential annotation information while removing visual clutter that distracts from the actual drawing features.
Solution Approach 2:
The patent performs preliminary detection and identification of lead lines and part numbers before final display or processing. By pre-identifying these elements through linear regression analysis and path tracing, the system can prepare multiple versions of the drawing (with or without annotations) in advance, allowing users to focus on desired information without manual editing.
3Measurement precision
If manual processing is used to derive information from patent drawings, then the information can be accurately extracted, but the process is slow and monotonous
Solution Approach 1:
The patent implements self-service automation where the system automatically detects, identifies, and processes lead lines and part numbers without human intervention. The linear regression analysis and path tracing algorithms enable the drawing information to extract and organize itself, maintaining high accuracy while dramatically increasing processing speed and eliminating manual labor.
Solution Approach 2:
The patent replaces manual mechanical processing with automated computational methods. Instead of human readers manually tracing and interpreting lead lines, the system uses linear regression analysis and digital image processing algorithms to automatically detect and interpret drawing annotations, achieving both speed and accuracy improvements.
4Ease of manufacture
If 2D drawings are converted directly to 3D models without removing extraneous information, then the conversion process is simpler, but the presence of lead lines and part numbers interferes with accurate model generation
Solution Approach 1:
The patent extracts and removes lead lines and part numbers before the 2D to 3D conversion process. By detecting and eliminating these extraneous elements that would interfere with model generation algorithms, the system prepares clean drawing data that enables accurate 3D model creation while maintaining the simplicity of the automated conversion process.
Data Source
AI summary
A method for removing extraneous information from a drawing of a device having one or more parts, the extraneous information including a lead line and a part number, the method including detecting a starting point of a lead line with respect to a bounding box of the part number and determining a general direction of the starting point with respect to the bounding box, wherein the starting point is set as a current point; performing linear regression analysis of dark pixels in an area covered by a mask centrally located at the current point to yield a linear regression line slope; determining a next point based on the slope and the general direction; and updating the current point with the next point and repeating the performing and determining steps until the lead line ends; and removing the lead line based on a path traced by the starting and current points.