Eye Gaze Tracking Using Neural Network and Kalman Filter
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing eye gaze tracking methods suffer from limited accuracy in real-world conditions, particularly without dedicated hardware and under varying eye appearance, illumination, head pose, and image quality, which restricts their usability and practicality in mainstream computing devices.
Innovation Solution
A computer-implemented method using an artificial neural network to construct a gaze vector from selected regions of interest, combined with a Kalman filter for precise point of gaze location on a screen, enhancing accuracy without relying on expensive hardware or user-dependent calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If dedicated hardware components (IR cameras, wearable hardware) are used for eye gaze tracking, then measurement precision is improved (accuracy of 0.5°), but device complexity and cost increase
Solution Approach 1:
The patent replaces dedicated hardware components (IR cameras, wearable devices) with a standard camera and software-based processing. The system uses a regular camera combined with an artificial neural network to detect eye landmarks and estimate gaze direction, eliminating the need for specialized mechanical or optical hardware while achieving competitive accuracy
Solution Approach 2:
The patent creates a software model (neural network) that learns to replicate the functionality of expensive hardware-based gaze tracking systems. By training the neural network on datasets from various eye tracking devices, the system copies the measurement capabilities of dedicated hardware using only standard camera inputs and computational algorithms
2Measurement precision
If user-dependent calibration procedures are implemented, then measurement precision is improved, but loss of time increases (calibration is time consuming)
Solution Approach 1:
The system performs automatic calibration without requiring user participation or time investment. The neural network is pre-trained on diverse datasets and automatically adapts to individual users through the calibration-free processing pipeline, eliminating the need for manual calibration procedures while maintaining accurate gaze estimation
Solution Approach 2:
The calibration data and model parameters are prepared in advance during the offline training phase. The neural network is pre-trained on extensive datasets containing various eye appearances, illuminations, and head poses, so that when the system is deployed, no additional calibration time is required from the user
3Device complexity
If standard cameras are used instead of dedicated hardware, then device complexity is reduced, but measurement precision deteriorates under varying conditions (eye appearance, illumination, head pose, image quality)
Solution Approach 1:
The neural network is trained to be invariant to changes in various parameters including illumination conditions, head pose angles, eye appearance variations, and image quality levels. By learning from diverse training data spanning these parameter variations, the system maintains stable and accurate gaze estimation across different operating conditions using only a standard camera
Solution Approach 2:
The system uses a universal neural network model that can handle multiple variations in input conditions (different illuminations, poses, eye appearances) through a single unified processing pipeline. This multi-functional approach allows the same hardware and software to perform accurate gaze tracking across diverse scenarios without requiring specialized components
Data Source
Figure 1
Figure 2
Figure 3a~3b
AI summary
The invention refers to a computer-implemented method for locating a point of gaze onto a screen (151). The method comprises the step (210) of initiating the acquiring of an image (300) and the steps (220, 230) of initiating the locating of a first face landmark location (301) and a second face landmark location (302) in the image (300). The method further comprises the step (240) of initiating the selecting of a region of interest (310) in the image (300), said selection being performed by using the aforementioned landmark locations (301,302). The method comprises the step (250) of initiating the constructing of a gaze vector, wherein the construction of the gaze vector is performed by means of an artificial neural network using the first region of interest (310) as input. Moreover, the method comprises the step (260) of initiating the locating the point of gaze onto the screen (151), wherein the locating of the point of gaze is performed by means of the gaze vector.