Document Image Obstruction Removal via Depth Data Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing methods fail to effectively remove obstructions from photographed images, such as PPTs and posters, especially when the photographer is not at an ideal location, leading to incomplete and unattractive corrections.
Innovation Solution
An image processing method and device that uses depth data to determine the presence of obstructions by establishing a reference plane and calculating differences in depth data between sampling points, allowing for the removal and restoration of obstructed information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Shape
If document correction is performed on photographed images, then the document can be corrected to a regular rectangle, but obstructions such as bodies or arms of spokespersons and heads of audience remain in the corrected image
Solution Approach 1:
The patent segments the image processing into multiple independent modules: obstruction detection module, depth data processing module, and document correction module. The obstruction detection module separately identifies obstructing objects using depth data, the depth data processing module calculates depth differences to determine obstruction locations, and the document correction module independently corrects the document shape. This segmentation allows each module to focus on specific tasks without interference, enabling both document shape correction and obstruction removal to be achieved simultaneously.
Solution Approach 2:
The patent performs obstruction detection and depth data processing before document correction. By preliminarily identifying obstructions and calculating depth differences to determine obstruction locations, the system prepares the image data in advance, allowing the subsequent document correction process to proceed without being blocked by obstructions. This preliminary action ensures that both obstruction removal and document correction can be effectively performed.
2Measurement precision
If depth data is used to detect obstructions, then obstruction detection accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent introduces depth difference data as an intermediary parameter to bridge the gap between raw depth data and obstruction detection. Instead of directly processing complex depth information, the system calculates depth differences between corresponding points in the photographed image and the reference plane, transforming the complex depth data into a simplified difference metric that directly indicates obstruction presence. This intermediary approach reduces processing complexity while maintaining detection accuracy.
Solution Approach 2:
The patent changes the parameter representation from absolute depth values to depth differences. By transforming the depth data into depth difference data (difference between photographed image depth and reference plane depth), the system simplifies the processing requirements. This parameter change allows for more straightforward comparison and threshold-based detection, reducing computational complexity while improving measurement precision for obstruction detection.
Data Source
AI summary
A method includes photographing a first image, where the first image comprises a document. A first area of the document is obscured by a first obstruction. The method further includes determining a location of the first area based on depth data. The method further includes photographing a second image. The method further includes restoring the obstructed information in the first area based on the second image. The method further includes displaying a third image, wherein the third image comprises the document and the first obstruction s removed from the first area of the document.