3D Face Shape Model Parameter Estimation via Depth Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing face image processing techniques face challenges in accurately estimating three-dimensional face shape model parameters due to ambiguity between coordinate values and vulnerability to environmental disturbances, leading to insufficient estimation accuracy.
Innovation Solution
A face image processing device and program that detect x and y coordinates in an image coordinate system, estimate z coordinates using deep learning, and convert these values to a camera coordinate system to derive accurate three-dimensional model parameters for a three-dimensional face shape model, incorporating a distance sensor and illumination unit to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If three-dimensional information is converted to two-dimensional information by projection conversion, then the processing can be performed in two-dimensional space, but ambiguity occurs between coordinate values and estimation error increases
Solution Approach 1:
The patent reverses the conventional approach by estimating the z-coordinate (depth) in three-dimensional space rather than projecting to two-dimensional space first. The z-coordinate estimation unit estimates depth information using feature points and three-dimensional face shape models, then converts to image coordinates only after estimation. This dimensional approach eliminates coordinate ambiguity while maintaining processing feasibility.
2Measurement precision
If three-dimensional sensor is used to acquire feature points, then depth information can be obtained directly, but the system becomes vulnerable to environmental disturbances such as sunlight
Solution Approach 1:
The patent replaces physical three-dimensional sensors with a computational estimation system. Instead of using hardware sensors that are vulnerable to sunlight and environmental factors, the system uses an estimation unit that calculates z-coordinates based on two-dimensional image features and three-dimensional face shape models. This substitution eliminates environmental susceptibility while maintaining depth estimation capability.
3Productivity
If optimization calculation is performed to minimize distance error in two-dimensional space, then computational efficiency is improved, but coordinate ambiguity between x-z and y-z values occurs
Solution Approach 1:
The patent performs optimization calculation in three-dimensional space by estimating the z-coordinate alongside x and y coordinates using feature points and projection relationships. This approach maintains computational efficiency through optimization while eliminating coordinate ambiguity, as the three-dimensional estimation provides unique depth information that resolves the ambiguity present in two-dimensional projection.
Data Source
AI summary
A face image processing device, includes: an image coordinate system coordinate value derivation unit detecting an x-coordinate value and a y-coordinate value in an image coordinate system at a feature point of an organ of a face of a person in an image, and estimating a z-coordinate value, so as to derive three-dimensional coordinate values in the image coordinate system; a camera coordinate system coordinate value derivation unit deriving three-dimensional coordinate values in a camera coordinate system from the three-dimensional coordinate values in the image coordinate system derived by the image coordinate system coordinate value derivation unit; and a parameter derivation unit applying the three-dimensional coordinate values in the camera coordinate system derived by the camera coordinate system coordinate value derivation unit to a predetermined three-dimensional face shape model to derive a model parameter of the three-dimensional face shape model in the camera coordinate system.