Geometry Generation from Handwritten Sketches via Pixel Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional computer-aided design (CAD) tools for generating geometry of objects, such as sheet metal parts, are tedious and time-consuming, requiring users to enter numerous dimensions manually, which can lead to inaccuracies and necessitate remodelling, and often require expertise in using complex graphic-user-interface (GUI) based tools.
Innovation Solution
A method using a computing device to generate geometry from a sketch drawn on a touch-sensitive display, where low-speed pixel points are joined with straight lines and processed to create a geometry with accurate bend angles and directions, allowing users to input additional dimensions for precise geometry creation, and enabling symmetry-based geometry generation to reduce user effort and dimension entry.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional CAD tools are used to generate geometry by entering cross-sectional dimensions, then the geometry can be generated with precise dimensions, but the process becomes tedious and time-consuming requiring extensive manual input
Solution Approach 1:
The patent replaces the manual mechanical input method (entering dimensions through GUI elements) with an optical input method (capturing images of handwritten sketches). The system uses image processing algorithms to automatically convert handwritten geometry into precise digital geometry models, eliminating the need for manual dimension entry while maintaining accuracy.
Solution Approach 2:
The system enables self-service by allowing users to input geometry through natural handwriting that the system automatically processes. The image processing and geometry generation happen automatically without requiring user expertise in CAD tools or manual dimension entry, making the process self-sufficient and eliminating the need for extensive manual intervention.
2Manufacturing precision
If conventional GUI-based modeling tools are used, then detailed geometry can be created, but the complexity of the interface requires user expertise and increases operation difficulty
Solution Approach 1:
The patent replaces the complex mechanical GUI interaction system with a simple optical capture system. Instead of requiring users to navigate menus, click buttons, and enter dimensions through a complex interface, the system simply captures an image of the user's handwritten sketch and automatically processes it into detailed geometry, dramatically simplifying the operation.
Solution Approach 2:
The system uses copying by capturing the visual representation of geometry (handwritten sketch) and converting it into a digital geometry model. This allows the system to replicate the user's intended geometry from a simple visual copy rather than requiring the user to manually construct it through complex GUI elements.
3Ease of manufacture
If progressive modeling is used where dimensions are entered portion-by-portion, then the geometry can be built incrementally, but the process becomes laborious requiring repeated manual input for each portion
Solution Approach 1:
The patent merges the entire geometry input process into a single operation. Instead of requiring separate input operations for each portion of the geometry, the system captures the complete handwritten sketch in one image and processes all portions simultaneously, combining multiple input steps into one unified action.
Solution Approach 2:
The patent extracts the essential geometry information directly from the handwritten sketch through image processing, removing the need for the complex progressive input process. The system extracts all necessary dimensional and geometric information from the single captured image, eliminating the need for repeated manual input operations.
Data Source
Figure 1
Figure 2(a)~2(b)
Figure 2(c)
AI summary
The present subject matter relates to generation of a geometry corresponding to a cross-sectional profile of an object. In an example method, a sketch corresponding to the cross-sectional profile of the object is received by a computing device (102). The sketch is drawn by moving a pointer over a display unit (104) of the computing device (102). Pixel points generated by the display unit (104) are obtained by the computing device (102). The pixel points correspond to the sketch. Low-speed pixel points, from amongst the pixel points, are joined with straight lines. The low-speed pixel points are the pixel points at which a speed of drawing of the sketch is below a predefined speed threshold. The straight lines are processed by the computing device (102) into the geometry of the object.