Privacy Information Occlusion in Picture Processing via Region Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing picture processing technologies inaccurately identify or miss privacy information, especially in cross-line text, leading to poor accuracy in privacy information identification and inconvenient user interactions.
Innovation Solution
A method combining region type detection and optical character recognition technologies to accurately identify and occlude privacy information in images, using a parallel processing mode to complement detection results and ensure accurate identification of entity content across lines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional privacy information identification methods are used, then the processing speed is fast, but the accuracy of privacy information identification is poor
Solution Approach 1:
The patent segments the privacy information identification process into two distinct modules: region type detection (identifying potential privacy regions) and character recognition (confirming actual privacy content). This segmentation allows each module to specialize in its task, improving overall accuracy while maintaining manageable system complexity through modular design.
Solution Approach 2:
The patent merges region type detection technology and optical character recognition technology into a unified processing framework. By combining these two complementary technologies, the system leverages the strengths of both: region type detection quickly identifies potential privacy areas, while character recognition accurately confirms the actual privacy content, thereby improving identification accuracy.
2Measurement precision
If traditional single-method detection is used, then the system complexity is low, but cross-line privacy information cannot be accurately identified
Solution Approach 1:
The patent introduces a new dimension to privacy information detection by combining spatial region analysis (region type detection) with textual character analysis (optical character recognition). This two-dimensional approach enables the system to accurately identify privacy information that spans multiple lines, as it can detect both the spatial distribution and the actual text content across different lines.
Solution Approach 2:
The patent uses region type detection as an intermediary step between image input and character recognition. This intermediary module identifies potential privacy regions and guides the character recognition process, improving the accuracy of cross-line privacy information identification while managing system complexity through staged processing.
3Measurement precision
If comprehensive privacy information identification is performed, then the accuracy improves, but the processing time increases
Solution Approach 1:
The patent performs preliminary region type detection before conducting character recognition. This preliminary action identifies potential privacy regions and narrows down the search scope, allowing the subsequent character recognition process to focus only on relevant areas. This staged approach improves identification accuracy while reducing overall processing time by avoiding unnecessary full-image analysis.
Data Source
AI summary
A picture processing method performed by an electronic device, includes: acquiring a target picture; performing region type detection and character recognition on the target picture by combining a target type detection technology and an optical character recognition technology, determining a target region to be occluded from the target picture based on a result of the region type detection and a result of the character recognition, and occluding the target region in the target picture.