Low-Angle Camera Facial Expression Tracking

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Wearable devices with cameras positioned close to or at low angles to the user cannot provide useful visual information, limiting their functionality in capturing facial expressions and eye movements for applications like avatar control.

Innovation Solution

Implementing low-angle cameras and eye-tracking cameras on wearable smart devices, combined with machine learning algorithms and depth sensors, to process images and derive valuable information for controlling avatars, even when cameras capture images nearly parallel to the user's face.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Volume of moving object

If cameras are positioned very close to and at low angles to the user, then the device can be made compact and wearable, but the images captured become nearly parallel to the user's face and lose useful visual information

Engineering Contradiction:
Improvedevice compactnessVSAvoidfacial expression information
Core Design Contradiction:
Volume of moving objectVSLoss of information

Solution Approach 1:

The patent introduces an intermediary processing system that captures low-angle images and uses machine learning algorithms to reconstruct meaningful facial expression data. The intermediary software layer transforms the limited low-angle camera data into useful facial expression information, resolving the contradiction between compact camera positioning and information quality.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system changes the parameter of image processing by applying machine learning models that learn to extract facial expression parameters from low-angle views. Instead of relying on traditional high-quality front-facing camera images, the system transforms the approach by learning to interpret distorted low-angle images and extract meaningful facial expression parameters.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If outward pointing cameras are used to capture useful views, then image quality improves, but the device cannot be positioned against the user's face for eye tracking and avatar control

Engineering Contradiction:
Improvefacial expression detection accuracyVSAvoideye tracking capability
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The patent merges the functions of outward-pointing cameras and inward-pointing eye-tracking cameras into a unified system. By combining data from both camera types and processing them together through machine learning, the system achieves both accurate facial expression detection and reliable eye tracking, resolving the contradiction between these two functional requirements.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The wearable device achieves multi-functionality by using a single integrated camera system that serves multiple purposes: capturing facial expressions for avatar control, tracking eye movements, and providing visual information. The machine learning processing enables the same hardware to perform multiple functions that would traditionally require separate specialized components.

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Device complexity

If low-angle cameras capture images parallel to the user's face, then the device structure is simplified, but deriving valuable information becomes significantly more difficult

Engineering Contradiction:
Improvecamera system structureVSAvoidfacial expression extraction
Core Design Contradiction:
Device complexityVSDifficulty of detecting and measuring

Solution Approach 1:

The patent replaces complex mechanical camera positioning with software-based intelligence. Instead of using multiple cameras positioned at various angles to capture facial expressions, the system uses a single low-angle camera combined with machine learning algorithms that simulate the computational work of multiple cameras, thereby simplifying the mechanical structure while maintaining detection capability.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system performs preliminary action by training machine learning models in advance with large datasets of facial expressions captured from various angles. This pre-trained knowledge enables the system to quickly and accurately extract facial expression information from low-angle images during actual use, reducing the computational difficulty of real-time detection.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3137938B1Facial expression tracking
Publication Date: 2018.01.31 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP3137938B1 patent drawingFigure 1
  • EP3137938B1 patent drawingFigure 2
  • EP3137938B1 patent drawingFigure 3

AI summary

The description relates to facial tracking. One example can include an orientation structure configured to position the wearable device relative to a user's face. The example can also include a camera secured by the orientation structure parallel to or at a low angle to the user's face to capture images across the user's face. The example can further include a processor configured to receive the images and to map the images to parameters associated with an avatar model.