Back to results
US 8,477,996 B2Granted
Method and device for finding and tracking pairs of eyes
AssigneeSeeReal Technologies S.A.
InventorEnrico Zschau
Life of this patent
Estimating the term…
Abstract
A method for finding and subsequently tracking the 3-D coordinates of a pair of eyes in at least one face, including receiving image data, which contains a sequence of at least one digital video signal of at least one image sensor, finding eyes or tracking previously found eyes in the image data, ascertaining the 3-D coordinates of the found or tracked eyes, associating the found or tracked eyes with a pair of eyes and providing the 3-D coordinates of the pair of eyes.
Claims26
A method for finding the 3D coordinates of a pair of eyes in at least one face, comprising the following steps:
- a) Reception of image data comprising a sequence of at least one digital video signal from at least one image sensor;
- b) Finding eyes in the image data;
- c) Determining the 3D coordinates of the found eyes;
- d) Assignment to the found eyes to form a pair of eyes;
- and
- e) Output of the 3D coordinates of the pair of eyes;
- where the process of finding eyes in the image data comprises the following steps:
- f) Identification of one or multiple faces in the image data;
- g) Determining the 3D coordinates of at least one identified face;
- h) Definition of a first search area on the identified face;
- and
- i) Finding at least one eye in the first search area;
- wherein the process of finding at least one eye in the first search area comprises the following steps:
- j) Computing an expected size of a part of the eye depending on the distance of the identified face from the image data providing image sensor;
- k) Computing a greyscale value as a threshold for a segmentation in the first search area depending on the distance of the identified face from the image data providing image sensor;
- l) Pre-processing the first search area aiming to improve the contrast;
- m) Segmentation of the first search area after the pre-processing;
- n) Computing one or multiple combined regions in the segmented first search area, where the combined region is a collection of adjacent pixels with at least approximately identical greyscale value;
- o) Determining a size of each computed combined region;
- and
- p) Comparing the expected size of the part of the eye with the determined size of the combined region, where the combined region represents a found eye if the determined size of the combined region at least approximately corresponds with the expected size of the part of the eye.
18 more elements
A method for tracking the 3D coordinates of a pair of eyes in at least one face, comprising the following steps:
- a) Reception of image data comprising a sequence of at least one digital video signal from at least one image sensor;
- b) tracking already found eyes in the image data;
- c) Determining the 3D coordinates of the tracked eyes;
- d) Assignment to the tracked eyes to form a pair of eyes;
- and
- e) Output of the 3D coordinates of the pair of eyes;
- wherein:
- the process of tracking already found eyes in the image data comprises the following steps:
- f) Definition of a search area on the identified face;
- and
- g) Tracking the eyes of the pair of eyes in the search area;
- and
- the process of tracking the eyes of the pair of eyes in the search area comprises the following steps:
- h) Computing an expected size of a part of the eye depending on the distance of the identified face from the image data providing image sensor;
- i) Determining a minimum greyscale value in the search area;
- j) Iteration over a greyscale value as a threshold in the search area, where the iteration is terminated when at least two eyes are detected, comprising the following steps:
- k) Computing the greyscale value as a threshold for a segmentation in the search area depending on the distance of the identified face from the image data providing image sensor, from the current iteration step, and from the minimum greyscale value;
- l) Segmentation of the search area;
- m) Computing one or multiple combined regions in the segmented search area, where the combined region is a collection of adjacent pixels with at least approximately identical greyscale value;
- n) Determining a size of each computed combined region;
- o) Comparing the expected size of the part of the eye with the determined size of the combined region, where the combined region represents a tracked eye if the determined size of the combined region at least approximately corresponds with the expected size of the part of the eye.
18 more elements
Dependencies are inferred from the claim text.