Head Feature Dimension Estimation From a Single Image
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for determining real-world dimensions and distances of head features require additional hardware, multiple images, or specific illumination, making them inconvenient for applications like spectacle lens fitting and virtual try-on technologies.
Innovation Solution
A computer-implemented method using probability distributions based on pixel dimensions and facial feature analysis to estimate real dimensions and distances without additional hardware, utilizing existing images and available data from extensive databases to refine the estimation process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods use additional hardware (reference objects, ultrasonic sensors, rotating platforms) to determine real dimensions and distances, then measurement precision is improved, but device complexity and ease of operation deteriorate
Solution Approach 1:
The system uses the head itself and its natural features (eyes, nose, ears, mouth) as reference objects for scaling. The head's own anatomical features serve as the measurement baseline, eliminating the need for external reference objects like credit cards or specialized markers.
Solution Approach 2:
The invention creates a digital model of the head from a single 2D image by copying and reconstructing three-dimensional geometric information. This digital twin allows virtual measurement and fitting without physical interaction with the subject.
2Measurement precision
If conventional methods use multiple images or specific illumination to determine real dimensions, then measurement precision is improved, but productivity and ease of operation deteriorate
Solution Approach 1:
The system pre-establishes a database of head feature dimensions and relationships that can be directly applied to new measurements. By having reference data ready in advance, the system can determine real dimensions from a single image without requiring multiple shots or complex illumination setups.
Solution Approach 2:
The invention transforms the problem from requiring multiple images with varying parameters (angles, illumination) to using a single image where computational parameters and probability distributions are adjusted to extract three-dimensional information through algorithmic processing.
3Measurement precision
If conventional methods require objects of known size for scaling, then measurement precision is improved, but ease of operation and adaptability deteriorate
Solution Approach 1:
The head's anatomical features serve as self-contained reference objects with known dimensional relationships. The system uses the natural geometry of facial features (eye separation, nose width, ear position) as built-in scaling references that are always present and require no external objects.
Solution Approach 2:
The system creates a universal measurement approach that works across different head sizes, ages, and ethnicities by using probability distributions derived from extensive databases. The same algorithmic framework adapts to various populations without requiring population-specific reference objects.
4Measurement precision
If conventional methods use extensive databases and probability distributions, then measurement precision is improved, but computational complexity increases
Solution Approach 1:
Extensive databases of head feature dimensions and probability distributions are pre-computed and stored during system initialization. This preliminary data preparation allows rapid querying and application during actual measurements without performing complex computations in real-time.
Solution Approach 2:
The system uses probability distributions that incorporate feedback from extensive databases of measured head dimensions. The computational model continuously refines estimates by comparing against statistical patterns from large datasets, improving accuracy through iterative probabilistic reasoning.
Data Source
AI summary
Computer implemented methods and devices for determining dimensions or distances of head features are provided. The method includes identifying a plurality of features in an image of a head of a person. A real dimension of at least one target feature of the plurality of features or a real distance between at least one target feature of the plurality features and a camera device used for capturing the image is estimated based on probability distributions for real dimensions of at least one feature of the plurality of features and a pixel dimension of the at least one feature of the plurality of features.