Face Distortion Correction Using Conformal Projection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies struggle with severe perspective deformation of faces at the edges of images captured by cameras with large fields of view, which negatively impacts user experience.
Innovation Solution
A method and apparatus for correcting face distortion, involving face detection, determination of the field of view range, and distortion correction using conformal projection, specifically targeting faces partially or fully outside the predetermined field of view range.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a camera with a large field of view is used to capture images, then the coverage area is increased, but severe perspective deformation occurs at the edges of the image
Solution Approach 1:
The patent applies different processing strategies to different regions of the image based on their field of view characteristics. Faces within the predetermined field of view range (central region) are preserved without correction, while faces outside this range (edge regions) undergo distortion correction processing. This local differentiation resolves the contradiction by maintaining high face shape accuracy in corrected regions while preserving the wide coverage area advantage.
Solution Approach 2:
The patent changes the distortion correction parameters based on the field of view position. By determining whether each face box is within the predetermined field of view range and applying selective correction, the system dynamically adjusts processing intensity. This parameter-based approach allows the system to correct edge distortion while maintaining overall image coverage.
2Manufacturing precision
If distortion correction is applied to all faces in the image, then face shape accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The patent extracts only the faces that require correction by identifying face boxes outside the predetermined field of view range. Instead of processing all faces uniformly, the system selectively applies distortion correction only to affected regions. This extraction approach significantly reduces processing time and computational resources while maintaining face shape accuracy where needed.
Solution Approach 2:
The patent applies partial action by correcting only the portion of faces that are actually distorted (those outside the field of view range). Rather than applying full correction to all faces, the system performs minimal necessary processing, thereby reducing time loss while achieving the required face shape accuracy for affected regions.
3Manufacturing precision
If faces at the edge of the image are corrected, then face quality in corrected regions is improved, but the quality of undistorted faces within the field of view may be degraded
Solution Approach 1:
The patent applies local quality by using different processing approaches for different image regions. Faces within the predetermined field of view range are preserved without correction to maintain their original high quality, while faces outside this range receive distortion correction. This regional differentiation ensures quality consistency across all faces by applying appropriate processing to each region.
Solution Approach 2:
The patent implements feedback by first determining the field of view range and then using this information to guide correction decisions. The system checks whether each face box is within the predetermined range and adjusts processing accordingly, ensuring that correction is applied only where needed and that overall quality consistency is maintained across the entire image.
Data Source
AI summary
The present disclosure provides a method for correcting face distortion, an apparatus for correcting face distortion, an electronic device, and a storage medium. The method includes: performing face detection on an obtained image to determine a position of each face box included in the image; determining whether each face box is within a predetermined field of view range based on the position of the face box; and performing distortion correction on a face in a first face box in response to at least a part of the first face box being not within the predetermined field of view range, to generate a corrected image.