Face Tracking Symbolic Point Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for monitoring vigilance and tracking facial movements in real-time are costly and inefficient due to the need for powerful image processing, making them unsuitable for mass-market applications like automobiles, where cost and precision are critical.
Innovation Solution
A method for initializing face tracking systems by identifying and tracking specific symbolic points on a face using low-cost imaging systems, involving steps such as detecting contrasted elements, selecting zones, determining natural points, calculating contrast gradients, and selecting points with high resemblance scores, to reduce processing complexity and maintain performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If powerful and expensive image processing means are used to achieve real-time face tracking, then measurement precision and tracking speed are improved, but device complexity and cost increase
Solution Approach 1:
The face image is divided into three specific zones (right eye zone, left eye zone, and mouth zone) based on anatomical landmarks. This segmentation allows the system to focus processing only on relevant regions rather than analyzing the entire image, reducing computational complexity while maintaining precision in identifying symbolic points such as eye corners and mouth corners.
Solution Approach 2:
The patent applies different processing strategies to different zones of the face image. Each zone is analyzed with specific algorithms tailored to its characteristics (e.g., detecting contrasted elements in eye zones versus mouth zone), allowing optimized processing that maintains high precision for each facial feature while reducing overall computational burden.
2Device complexity
If low-cost imaging systems are used to reduce device complexity and cost, then device complexity is reduced, but measurement precision and tracking capability deteriorate
Solution Approach 1:
The system performs preliminary detection of contrasted elements (such as eyes, nostrils, and mouth) to establish zones of interest before conducting detailed symbolic point identification. This preliminary action prepares the data structure and focuses subsequent processing on specific regions, enabling low-cost systems to achieve adequate precision by avoiding unnecessary processing of irrelevant image areas.
Solution Approach 2:
The patent implements a multi-stage processing approach where only certain critical features are detected with high precision (symbolic points at contrast intersections), while other areas receive minimal or no processing. This partial action strategy ensures that the most important tracking points are identified accurately even with limited computational resources.
Data Source
AI summary
A method for identifying symbolic points on the image of a face including images of a right eye, left eye and mouth, includes:detecting and identifying elements with strong contrasts such as the irises, nostrils, or mouth;selecting zones of the image with respect to the elements with strong contrasts including a priori two sought-after symbolic points interrelated by a morphological criterion;searching within the zones for natural points through the convergence of lines of the image and, for each natural point, determining a signature, determining a score with respect to pre-established signatures and selecting the natural points having a score above a threshold value;in each zone, identifying pairs of natural points and determining for each identified pair a score with respect to pairs of standard symbolic points and selecting the pair of natural points having the best score as symbolic points of the zone.