Casting Mold Texture Expansion With Seamless Neural Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating mold textures for casting molds struggle to produce seamless, large-scale textures that mimic organic structures like skin or snake skin, often resulting in visible artifacts and high production costs due to the need for scanning or hand-crafting.
Innovation Solution
A method using a generative neural network to extend a seed texture to a larger size, allowing for the creation of seamless and high-quality mold textures that can be scaled arbitrarily, while maintaining the look and feel of the seed texture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a seed texture is scanned or hand-crafted to create a mold texture, then the texture quality and realism are improved, but the production cost and time increase significantly
Solution Approach 1:
The patent uses a seed texture as a template and automatically generates multiple variations through neural network processing. The system copies the essential characteristics of the seed texture and produces seamless mold textures without requiring manual scanning or hand-crafting of each texture, significantly reducing production cost while maintaining quality
Solution Approach 2:
The patent applies neural network-based parameter transformations to the seed texture, adjusting texture parameters such as scale, orientation, and pattern variations automatically. This allows high-quality texture generation through computational parameter changes rather than manual processes, reducing both cost and time
2Area of stationary object
If the mold texture size is increased to cover larger mold surfaces, then the coverage area is improved, but visible repetitions and artifacts appear in the texture
Solution Approach 1:
The patent transitions from 2D texture tiling to 3D seamless texture generation using neural networks. By processing the texture in multiple dimensions and considering spatial relationships across the entire mold surface, the system generates coherent textures that maintain realism at any scale without visible repetitions or artifacts
Solution Approach 2:
The patent uses dynamic neural network processing to adaptively generate texture patterns based on the required mold surface area. The system dynamically adjusts texture parameters and patterns to ensure seamless coverage, rather than statically tiling fixed-size textures that would show repetitions at larger scales
3Ease of operation
If a fixed-size seed texture is used as input, then the processing simplicity is improved, but the output texture size is limited by the neural network structure
Solution Approach 1:
The patent creates a universal neural network system that can process a single seed texture and generate outputs at multiple size scales. The network is designed to be scalable and adaptable, allowing the same input to produce various output dimensions without requiring separate processing for each size, thus maintaining simplicity while achieving versatility
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method for generating a mold texture (3) for a casting mold (1). It is provided that the mold texture (3) is generated from a seed texture (4), the mold texture (3) having a larger texture size in at least one dimension than the seed texture (4), wherein the seed texture (4) is provided as an input texture for a generative neural network (7) with a plurality of neural network parameters determined during training of the generative neural network (7) and the generative neural network (7) is used to extend the seed texture (4) to the texture size of the mold texture (3). The invention further relates to a device for generating a mold texture (3) for a casting mold (2), a computer program and a computer-readable medium.