AI-Generated Image Vectorization for Quality-Preserving Scaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Images generated by generative adversarial networks (GANs) are typically in raster format, which are not suitable for size changes without causing image quality deterioration, such as blurring at enlarged portions.
Innovation Solution
Convert raster images to vector format while considering conversion errors, using a discriminator to ensure authenticity and store only images that maintain original quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If images are generated using generative adversarial networks (GANs), then high image quality is achieved, but the images are in raster format and not suitable for size changes without quality deterioration
Solution Approach 1:
The patent converts the image format parameter from raster to vector, fundamentally changing how the image stores visual information. This parameter change enables the image to maintain quality during size changes while preserving the high quality characteristics of GAN-generated images
Solution Approach 2:
The patent creates a vector format copy of the GAN-generated raster image. This vector copy can be scaled to any size without quality loss, solving the adaptability problem while maintaining the original high quality through the conversion process that preserves visual fidelity
2Length of stationary object
If raster images are greatly enlarged, then image size increases, but blurring occurs and image quality deteriorates
Solution Approach 1:
Converting from raster to vector format changes the fundamental parameter of image representation. Vector images use mathematical descriptions of shapes and paths rather than fixed pixel grids, allowing infinite scaling without blurring or quality deterioration
Solution Approach 2:
The patent transitions from a discrete pixel-based dimension to a continuous mathematical dimension. Vector images can be rendered at any resolution by evaluating the mathematical functions, providing scalability across different size dimensions without quality loss
Data Source
AI summary
In an information processing method executed by an information processing device having a processor, the processor executes generating a first image in raster format using a generative adversarial network that includes a generator and a discriminator, converting the first image into a second image in vector format, inputting a third image obtained by converting the second image into raster format, or the second image, to the discriminator, and acquiring discrimination results, and storing the second image in a storage unit, if the discrimination results are true, or if the discrimination results are the same as discrimination results by the discriminator regarding the first image.