Back to results
US 11,113,842 B2Granted
Method and apparatus with gaze estimation
AssigneeSamsung Electronics Co., Ltd.
InventorsXiabing Liu, Hui Zhang, Jae-Joon Han, Changkyu Choi, Tianchu Guo
Life of this patent
Estimating the term…
Abstract
A gaze estimation method and apparatus is disclosed. The gaze estimation method includes obtaining an image including an eye region of a user, extracting, from the obtained image, a first feature of data, obtaining a second feature of data used for calibration of a neural network model, and estimating a gaze of the user using the first feature and the second feature.
Claims18
A processor-implemented gaze estimation method, comprising:
- obtaining an image including an eye region of a user;
- extracting, from the obtained image, a first feature of data;
- obtaining a second feature of data used for calibration of a neural network model;
- estimating a gaze of the user using the first feature and the second feature;
- and
- estimating a position of a gaze point in a gaze area of the user from the first feature and the second feature using the neural network model, including:
- calculating, using the neural network model, a probability that the gaze point is included in each of subareas into which the gaze area is divided, based on a calculated feature difference between the first feature and the second feature;
- and
- determining, to be the position of the gaze point, a center position of a subarea having a greatest probability among probabilities calculated for the subareas.
6 more elements
A processor-implemented gaze estimation method, comprising:
- obtaining an image including an eye region of a user;
- extracting, from the obtained image, a first feature of data;
- obtaining a second feature of data used for calibration of a neural network model;
- estimating a gaze of the user using the first feature and the second feature;
- displaying a calibration point;
- obtaining a calibration image which is obtained by capturing the obtained image of the user when the user gazes at the displayed calibration point;
- and
- performing the calibration on the neural network model using the obtained calibration image,
- wherein the obtaining of the calibration image comprises:
- when a gesture performed by the user in response to the calibration point is received, determining a distance between an operation point corresponding to the gesture and the calibration point;
- and
- in response to the determined distance being less than a threshold value, determining the obtained calibration image to be an image to be used to perform the calibration,
- wherein the performing of the calibration on the neural network model comprises:
- performing the calibration on the neural network model using the calibration image.
11 more elements
A gaze estimation apparatus comprising:
- one or more processors configured to:
- obtain an image including an eye region of a user;
- control the gaze estimation apparatus to extract a first feature of data;
- obtain a second feature of data used for calibration of a neural network model;
- estimate a gaze of the user using the extracted first feature and the second feature;
- and
- estimate a position of a gaze point in a gaze area of the user from the first feature and the second feature using the neural network model, including:
- calculating, using the neural network model, a probability that the gaze point is included in each of subareas into which the gaze area is divided, based on a calculated feature difference between the first feature and the second feature;
- and
- determining, to be the position of the gaze point, a center position of a subarea having a greatest probability among probabilities calculated for the subareas.
7 more elements
Dependencies are inferred from the claim text.