Spectacle Frame Edge Detection Using 3D Model Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the contours of spectacle frames in images require pre-recorded tracer data, which is inconvenient for opticians as it necessitates additional steps and storage, and typically only provides two-dimensional data when three-dimensional data is needed.
Innovation Solution
A computer-implemented method that automatically detects spectacle frame edges in multiple views using a parametric three-dimensional model, eliminating the need for pre-recorded tracer data by calibrating images and optimizing geometric parameters for accurate centration, allowing for three-dimensional representation and simplification of edge detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pre-recorded tracer data are used for contour detection, then contour finding can be performed, but additional storage requirements and workflow steps are necessary
Solution Approach 1:
The system performs self-calibration by automatically determining the relationship between multiple camera views and the 3D model without requiring pre-recorded tracer data. The calibration process uses the spectacle frame itself as the reference object, eliminating the need for external tracer data storage and processing.
Solution Approach 2:
The patent extracts the calibration information directly from the images of the spectacle frame worn by the customer, removing the dependency on separate tracer data records. The essential geometric relationships are derived from the actual spectacle being measured rather than from pre-stored reference data.
2Device complexity
If tracer data are determined after the customer leaves, then storage requirements are reduced, but centration cannot be performed before the customer departs
Solution Approach 1:
The system performs calibration and centration calculations immediately during the customer's visit by processing images captured in real-time. The 3D model is calibrated and centration parameters are computed on-the-spot, enabling same-day service completion without requiring post-visit data processing.
Solution Approach 2:
The patent replaces the traditional mechanical workflow of physical tracer card placement and manual data transfer with an automated digital imaging and processing system. Multiple calibrated cameras capture images that are automatically processed to generate centration data instantly.
3Ease of manufacture
If two-dimensional tracer data are used, then data acquisition is simple, but three-dimensional information is not available for accurate centration
Solution Approach 1:
The patent transitions from 2D tracer data to 3D spatial information by using multiple calibrated cameras to capture images from different viewpoints. These 2D images are processed through 3D coordinate transformations to reconstruct the spatial relationships of the spectacle frame and determine accurate 3D centration parameters.
Solution Approach 2:
The 3D model serves multiple functions: it acts as both the reference object for calibration and the target object for centration measurement. The same parametric 3D model of the spectacle frame is used to establish camera geometry and to calculate lens positioning parameters, eliminating the need for separate 2D tracer data.
Data Source
AI summary
A computer-implemented method for determining a representation of a rim of a spectacles frame or a representation of the edges of the spectacle lenses is disclosed, wherein at least two calibrated images taken from different viewing angles of a head a subject wearing the spectacles frame or the spectacles are provided, and wherein data for at least portions of the rims of the spectacles frame or the edges of the lenses are detected in each image. Further, a three-dimensional model of the spectacles frame or the spectacles is provided, based on geometric parameters, and the geometric parameters are optimised to adapt the model to the detected edges.