AR Text Blending via Background Interpolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing augmented reality systems face challenges in translating text on mobile devices without introducing undesirable artifacts in the background, due to limited computing power and complexity in re-rendering processes.
Innovation Solution
The method employs interpolation techniques, including linear and non-linear interpolation, to regenerate the background and reduce artifacts by isolating symbols using a detection module, path finder, background interpolator, translator, and pixel blender within a mobile device's blending system, optimizing paths to minimize noise and computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If re-rendering of the background is performed to replace text with translated text, then translation accuracy is improved, but artifacts are introduced in the background and computational complexity increases
Solution Approach 1:
The patent segments the image into text regions and background regions, allowing independent processing of each region. The text regions are identified and isolated, then replaced with translated text, while the background regions are separately processed through interpolation to remove artifacts caused by text replacement.
Solution Approach 2:
The patent extracts the text regions from the original image using detection modules, separates them from the background, and processes them independently. This extraction allows the background to be regenerated without text interference, eliminating artifacts while maintaining translation accuracy.
2Measurement precision
If re-rendering of the background is performed to replace text with translated text, then translation accuracy is improved, but computational complexity and power consumption increase
Solution Approach 1:
The patent applies different processing methods to different regions of the image. High-computation text detection and translation are applied only to text regions, while simpler interpolation methods are applied to background regions. This local differentiation reduces overall computational complexity while maintaining translation accuracy where needed.
Solution Approach 2:
The patent performs preliminary detection of text regions before full translation processing. By identifying text boundaries and regions in advance, the system can limit subsequent high-computation translation and background regeneration operations to only the necessary areas, reducing overall computational complexity.
3Measurement precision
If re-rendering of the background is performed to replace text with translated text, then translation accuracy is improved, but power consumption increases
Solution Approach 1:
The patent applies power-efficient processing by concentrating high-computation operations only on text regions rather than processing the entire image. The background regions use simpler, less power-intensive interpolation methods, significantly reducing overall power consumption while maintaining translation accuracy.
Solution Approach 2:
The patent performs preliminary text detection to identify only the regions requiring translation processing. This preliminary step allows the system to avoid unnecessary power consumption on regions that don't require translation, optimizing energy usage while maintaining translation accuracy for relevant content.
Data Source
Figure 1
Figure 2~2(B)
Figure 3
AI summary
The use of optical character recognition (OCR) in mobile devices is becoming prevalent with the increasing use of mobile devices. One important application for OCR in mobile devices is recognizing and translating the text to a language understandable by the user. Techniques are provided for replacing symbols in an image, while reducing the artifacts as a result of re-rendering of the background image.