Tile-Based Preview Image Generation for Virtual Spaces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high computational load of image stitching technologies, particularly in generating high-resolution images, results in significant resource consumption and slow preview times, even with acceleration techniques like GPU acceleration, due to the need for downsampled low-resolution images for speed.
Innovation Solution
A processor-implemented method that determines identification information of tile images within a visible region of a preview interface, updates tile images based on zoom levels and user gestures, and generates preview images by stitching partial images from a database, reducing the need for full image stitching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a downsampled low-resolution stitched image is generated to ensure preview speed, then the processing time is reduced, but the image details are not displayed properly
Solution Approach 1:
The patent divides the large virtual image into multiple tile images arranged in a grid pattern. Each tile represents a portion of the overall stitched image. This segmentation allows the system to process and display only the necessary tiles for the current visible region, rather than processing the entire high-resolution image, thus achieving fast preview while maintaining detail quality in the displayed regions.
Solution Approach 2:
The patent introduces a pyramid structure with multiple zoom levels, adding a dimensional aspect to image storage. Each level contains downsampled versions of the virtual image at different resolutions. This dimensional approach allows the system to quickly select appropriate resolution levels for preview, achieving fast display speed while preserving the ability to access high-resolution details when needed.
2Manufacturing precision
If an image fusion algorithm is used to stitch a high-resolution image, then the image quality is improved, but the computational load and resource consumption increase significantly
Solution Approach 1:
The patent pre-processes the virtual image by dividing it into multiple tiles and creating a pyramid structure with downsampled versions at different zoom levels before the actual stitching operation. This preliminary action stores the image data in an optimized format, allowing the fusion algorithm to work with pre-segmented tiles rather than the entire high-resolution image, thus reducing computational load while maintaining quality.
Solution Approach 2:
The patent applies different processing qualities to different regions and levels of the image pyramid. High-resolution details are maintained in the original virtual image and specific tiles, while lower-resolution versions are used for overview and navigation. This local quality approach ensures high image quality where needed while reducing overall computational resource consumption.
3Loss of information
If the entire virtual image is stitched to generate a preview, then the complete image is available, but the processing time is too long for interactive use
Solution Approach 1:
The patent extracts only the necessary portion of the virtual image for preview by identifying which tiles fall within the current visible region defined by the preview interface. Instead of processing the entire virtual image, the system extracts and processes only the relevant tiles, dramatically reducing processing time while maintaining the completeness of the displayed content within the visible region.
Solution Approach 2:
The patent implements a dynamic tile selection mechanism that adapts to user interactions such as panning and zooming. As the user navigates the preview interface, the system dynamically determines which tiles are currently visible and processes only those tiles. This dynamic approach ensures that the preview remains complete for the current view while minimizing processing time by excluding invisible regions.
Data Source
AI summary
A processor-implemented method includes: determining whether a plurality of tile sets of different zoom levels comprises a tile image, based on either one or both of a position of the tile image in a virtual image and a size of the tile image, wherein tile image corresponds to a visible region covered by a preview interface in the virtual image; updating the tile image based on whether the plurality of tile sets comprises the tile image; and generating a preview image corresponding to the tile image by stitching the updated tile image.