Raster Image Layer Decomposition for Accurate Design Element Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to accurately and flexibly segment design elements in digital images with multiple layers at different depths, often failing to distinguish between design elements and objects within a design element and not recognizing layering information, which complicates downstream operations.
Innovation Solution
An image decomposing system utilizes segmentation neural networks to decompose digital images into constituent layers, generating segmentation masks and classifying design elements, and inpaints regions with gaps, providing the inpainted layers for display via a graphical user interface.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing systems are used to segment digital images, then segmentation can be performed, but accuracy in distinguishing design elements and recognizing layering information deteriorates
Solution Approach 1:
The patent applies segmentation by dividing the digital image into multiple discrete layers, where each layer contains specific design elements. The system uses neural networks to segment the image into foreground layers (containing design elements) and background layers, and further segments each layer into individual design element masks. This multi-level segmentation approach enables accurate distinction between design elements and recognition of layering information.
2Ease of operation
If raster format images are processed, then image data can be accessed, but layer information retrievability deteriorates
Solution Approach 1:
The patent introduces an intermediary processing system that analyzes raster images to infer and reconstruct layer information. The neural network acts as an intermediary that processes the flat raster image data and generates layered segmentation masks, effectively recovering layer information that is not directly accessible in raster format. This intermediary process enables both image data accessibility and layer information retrieval.
3Adaptability or versatility
If design elements are segmented from images, then editing flexibility improves, but downstream operation complexity increases
Solution Approach 1:
The patent extracts design elements from images by generating separate segmentation masks for each element. The system takes out individual design elements (such as text, shapes, and images) from the composite image and provides them as distinct, editable components. This extraction process improves editing flexibility while reducing downstream operation complexity by providing cleanly separated elements ready for use.
Data Source
AI summary
The present disclosure relates to systems, non-transitory computer-readable media, and methods for decomposing a raster design into constituent elements. In particular, the disclosed systems determine, utilizing a plurality of segmentation neural networks, a set of layers corresponding to different depths of a digital image, each layer comprising non-overlapping design elements. In addition, the disclosed systems generate, utilizing the plurality of segmentation neural networks, segmentation masks for the digital image by decomposing the digital image into the design elements within the set of layers. Moreover, the disclosed systems provide, for display via a graphical user interface of a client device, the digital image with the design elements within the set of layers according to the segmentation masks.