Hierarchical Eye Tracking for Multi-Observer Real-Time Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing contactless eye detection and tracking systems struggle to accurately and efficiently track eye positions in real-time, especially with quick and large movements in all three dimensions, and fail to handle high-resolution cameras and multiple observers effectively, leading to delays and errors.
Innovation Solution
A hierarchical method using separate instances for face detection, eye region detection, and eye tracking, executed in parallel on dedicated computing units, which reduces the search region and data volume progressively, allowing for efficient and accurate real-time processing of eye positions across a large depth range.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If conventional eye detection methods are used, then basic eye position detection is achieved, but real-time tracking of abrupt movements in large depth ranges cannot be accomplished
Solution Approach 1:
The system segments the detection task into multiple hierarchical instances: face finder instances detect faces in the entire video frame, eye finder instances detect eyes within detected face regions, and eye tracker instances track eye positions within detected eye regions. This segmentation allows parallel processing of different regions at appropriate levels of detail, enabling real-time tracking of abrupt movements while maintaining detection accuracy through specialized processing at each hierarchical level.
2Area of stationary object
If high-resolution cameras are used to detect eyes at large distances, then detection range is improved, but data volume and processing complexity increase significantly
Solution Approach 1:
The system applies local quality by processing different regions of the video frame at different levels of detail. Face finder instances process the entire frame at low resolution to locate faces. Eye finder instances then process only the detected face regions at medium resolution to locate eyes. Eye tracker instances process only the detected eye regions at high resolution to precisely track eye positions. This localized quality adjustment reduces overall processing complexity while maintaining detection range capability.
3Adaptability or versatility
If multiple observers are tracked simultaneously, then system versatility is improved, but processing time and computational load increase
Solution Approach 1:
The system segments the multi-observer tracking task by creating independent face finder instances for different regions of the video frame. Each face finder instance independently detects faces in its assigned region, and subsequent eye finder and eye tracker instances process only the relevant regions. This segmentation enables parallel processing of multiple observers simultaneously, maintaining real-time performance while achieving multi-observer versatility.
4Reliability
If the entire video frame is processed for eye detection, then detection completeness is improved, but processing speed decreases
Solution Approach 1:
The system performs preliminary action by first detecting faces in the entire video frame using face finder instances. Once faces are detected, eye finder instances are activated only within the detected face regions to locate eyes. Finally, eye tracker instances process only the detected eye regions to precisely track eye positions. This preliminary detection and region restriction approach ensures detection completeness while dramatically improving processing efficiency by avoiding unnecessary processing of irrelevant frame regions.
Data Source
AI summary
The invention relates to a method and to a circuit arrangement for recognising and for tracking, in a contact-free manner, eye positions of several users in real time. The input data comprises a sequence of digital video frames. Said method comprises the following steps: combining a face-finder-instance which is used to examine faces, an eye-finder-instance which is used to examine eye areas, and an eye-tracker-instance which is used to recognise and track eye reference points. The aim of the invention is to convert the eye positions within a hierarchical outlet of the instance to the target, which successively restricts the dataset, which is to be processed, emerging from the dataset of the entire video frame (VF) in order to form a face target area (GZ) and subsequently an eye target area (AZ). Also, an instance or a group of instances, which run in a parallel manner, are carried out, respectively, on a calculating unit thereof.