Gaze Locking Detection Using Eye Region Feature Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional approaches lack mechanisms for directly sensing eye contact, which is crucial for human-object interactions and social interactions, making it difficult to determine whether a user is looking at a device or not.
Innovation Solution
A system and method for detecting gaze locking by receiving an input image, locating a pair of eyes, generating a coordinate frame, identifying an eye region, and determining gaze locking based on a feature vector generated from pixel values using a hardware processor and suitable classifiers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional approaches are used for sensing user attention, then the system lacks direct eye contact detection capability, but implementing such capability requires complex image processing and feature analysis
Solution Approach 1:
The patent segments the face detection process into distinct stages: face detection, eye pair localization, coordinate frame generation, eye region identification, and feature vector extraction. Each stage processes a specific aspect of the image, breaking down the complex task of eye contact detection into manageable segments that can be handled by specialized algorithms.
Solution Approach 2:
The patent extracts only the relevant features from the image data needed for eye contact detection. Instead of analyzing the entire image, the system extracts pixel values specifically from the eye region, generates feature vectors from these extracted values, and uses only these features for classification. This extraction principle reduces data processing requirements while maintaining detection accuracy.
2Measurement precision
If detailed feature vectors are generated from eye region pixels to improve detection accuracy, then the processing time and computational resources increase
Solution Approach 1:
The patent extracts only the necessary pixel values from the eye region rather than processing the entire image. By limiting the analysis to the specific eye region and extracting only relevant pixel features, the system reduces the computational burden while maintaining sufficient accuracy for gaze locking detection.
Solution Approach 2:
The patent transforms raw pixel values into feature vectors that capture the essential characteristics of the eye region in a compressed format. This parameter transformation converts high-dimensional pixel data into a more manageable feature space that retains the critical information needed for detection while reducing computational requirements for classification.
Data Source
AI summary
Methods, systems, and media for detecting gaze locking are provided. In some embodiments, methods for gaze locking are provided, the methods comprising: receiving an input image including a face; locating a pair of eyes in the face of the input image; generating a coordinate frame based on the pair of eyes; identifying an eye region in the coordinate frame; generating, using a hardware processor, a feature vector based on values of pixels in the eye region; and determining whether the face is gaze locking based on the feature vector.