3D Eyewear Recommendation Using Facial Mesh and Gaze Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional eyewear recommendation systems rely heavily on subjective assessment and manual measurements, failing to account for the intricate nuances of facial structures, eye dynamics, and individual preferences, leading to suboptimal outcomes.
Innovation Solution
A system utilizing deep learning, computer vision, and behavioral analysis to provide personalized eyewear recommendations by integrating depth estimation from RGB images, facial mesh alignment, and gaze tracking, enabling precise frame and lens suggestions that complement facial features and natural visual behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional subjective assessment and manual measurements are used for eyewear recommendation, then the system is simple to operate, but the recommendation precision and personalization are insufficient
Solution Approach 1:
The patent replaces manual mechanical measurements with automated computer vision and deep learning systems. The system uses RGB-D cameras and neural networks to automatically capture and analyze facial geometry, substituting the mechanical ruler-and-eye method with digital image processing and 3D reconstruction algorithms, thereby achieving millimeter-level precision without manual intervention
Solution Approach 2:
The patent creates a digital 3D copy of the user's facial structure through photogrammetry and depth mapping. This virtual facial model serves as a precise replica that can be measured, analyzed, and used for virtual try-on, eliminating the need for physical measurements while maintaining high accuracy
2Adaptability or versatility
If advanced data analytics and machine learning algorithms are implemented, then the recommendation personalization improves, but the computational resources and processing time increase
Solution Approach 1:
The patent pre-trains deep learning models on large datasets of facial geometries and eyewear fittings before deployment. This preliminary training allows the system to make rapid, energy-efficient predictions during actual use, as the complex pattern recognition has already been learned during the offline training phase
Solution Approach 2:
The patent divides the complex recommendation task into separate modular components: facial geometry extraction, feature point detection, frame compatibility analysis, and virtual try-on rendering. Each module processes specific aspects independently, reducing overall computational complexity and energy requirements compared to a monolithic approach
3Reliability
If virtual try-on technology with accurate facial mapping is used, then the recommendation accuracy improves, but the device complexity and implementation difficulty increase
Solution Approach 1:
The patent introduces a virtual 3D facial model as an intermediary between the physical face and the eyewear recommendation. This digital twin serves as a mediator that simplifies the interaction, allowing frames to be virtually tried on the digital model rather than requiring complex real-time physical overlay and adjustment mechanisms
Data Source
AI summary
System and methods for recommending eyewear to a user are disclosed. The method may include receiving one or more images of a face of the user, extracting spatial information from one or more images using a deep learning model to estimate depth and contours of one or more facial features, reconstructing a three-dimensional facial model based on the extracted spatial information, wherein the reconstruction includes aligning a face mesh with a reference object of known size or an alternative scaling mechanism, tracking eye movement and analyzing gaze fixation points to determine natural viewing behavior, and/or generating an eyewear recommendation based optionally on the three-dimensional facial model and/or the analyzed viewing behaviors, wherein the recommended eyewear is selected to optimize fit and/or visual alignment and/or user explicit or implicit preferences.