Gaze Direction Estimation Using Mobile Device Camera
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driver monitoring systems, especially those using mobile devices, lack accuracy in detecting driver gaze direction, particularly for predicting drowsiness, inattention, and distractions that can lead to unsafe driving conditions, and are not cost-effective or reliable.
Innovation Solution
A method and system that uses a mobile device-mounted camera to capture video of the driver, detect facial regions, compute feature descriptors, and apply a gaze classifier to identify the driver's gaze direction, trained using both stationary and moving vehicle data to improve accuracy and real-time execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If specialized built-in vehicle monitoring systems are used, then measurement precision of driver gaze direction is improved, but device complexity and cost increase
Solution Approach 1:
The patent uses a mobile device (smartphone/tablet) with a camera to capture driver images, copying the function of specialized built-in monitoring systems. The mobile device runs a gaze estimation application that processes images to determine driver gaze direction, avoiding the need for complex dedicated hardware while achieving comparable measurement precision through software-based computer vision algorithms
Solution Approach 2:
The patent replaces complex mechanical/optical monitoring hardware with a mobile computing platform that uses image processing and machine learning algorithms. The gaze estimation system uses a trained neural network model that processes camera images to predict gaze direction, substituting physical monitoring mechanisms with computational methods
2Device complexity
If mobile monitoring devices are used, then device complexity is reduced, but measurement precision of driver gaze direction deteriorates
Solution Approach 1:
The patent performs preliminary training of the gaze estimation model using a training data set before actual use. The system is calibrated with sample images and corresponding gaze annotations to train a neural network model that can accurately predict gaze direction. This preliminary training action enables the mobile device to achieve high measurement precision without requiring complex real-time processing hardware
Solution Approach 2:
The mobile device uses its own built-in camera and processing capabilities to perform gaze estimation independently. The system leverages the mobile device's existing sensors and computational resources rather than requiring external specialized equipment, enabling accurate gaze detection through self-contained hardware and software
3Productivity
If real-time gaze monitoring is implemented, then productivity of safety monitoring is improved, but use of energy increases
Solution Approach 1:
The patent implements periodic image capture rather than continuous monitoring. The mobile device captures images at specific intervals or when certain conditions are met (such as driver head movement detected by accelerometers), processing images only when necessary. This periodic action maintains real-time monitoring capability while significantly reducing energy consumption compared to continuous video streaming and processing
Data Source
AI summary
Methods and systems for continuously monitoring the gaze direction of a driver of a vehicle over time. Video is received, which is captured by a camera associated with, for example, a mobile device within a vehicle, the camera and/or mobile device mounted facing the driver of the vehicle. Frames can then be extracted from the video. A facial region can then be detected, which corresponds to the face of the driver within the extracted frames. Features descriptors can then be computed from the facial region. A gaze classifier derived from the vehicle, the driver, and the camera can then be applied, wherein the gaze classifier receives the feature descriptors as inputs and outputs at least one label corresponding to one or more predefined finite number of gaze classes to identify the gaze direction of the driver of the vehicle.