# US 2015/0104081 A1 · Methods and systems for determining user liveness

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2015/0104081 A1
- Document ID: US20150104081A1
- Status: Published application
- Kind: A1
- Application number: 14/053,384
- Filed: 2013-10-14
- Published: 2015-04-16
- Earliest claimed: 2013-10-14
- Assignees: Daon Holdings Limited
- Inventors: Mircea IONITA
- Web page: https://patentgrep.com/patent/US20150104081A1

## CPC

- G06V 40/45: {Detection of the body part being alive}
- G06T 7/60: Analysis of geometric attributes
- G06T 7/73: using feature-based methods
- G06V 10/446: {using Haar-like filters, e.g. using integral image techniques}
- G06V 40/19: Sensors therefor
- G06V 40/193: {Preprocessing; Feature extraction}
- G06V 40/197: {Matching; Classification}
- G06T 2207/30196: Human being; Person
- G06V 40/58: {Solutions for unknown imposter distribution}
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

A method determining user liveness is provided that includes calculating, by a device, eye openness measures for a frame included in captured authentication data, and storing the eye openness measures in a buffer of the device. Moreover the method includes calculating confidence scores from the eye openness measures stored in the buffer, and detecting an eye blink when a maximum confidence score is greater than a threshold score.

## Claims

### Claim 1

A method for determining user liveness comprising:
calculating, by a device, eye openness measures for a frame included in captured authentication data, and storing the eye openness measures in a buffer of the device;
calculating confidence scores from the eye openness measures stored in the buffer; and
detecting an eye blink when a maximum confidence score is greater than a threshold score.

### Claim 2 (depends on 1)

A method for determining user liveness in accordance with claim 1, said calculating eye openness measures step comprising:
positioning detection windows included in a table, on a region of interest included in the frame, at coordinates included in the table;
calculating a feature value for each window position; and
calculating the eye openness measures using the feature values.

### Claim 3 (depends on 2)

A method for determining user liveness in accordance with claim 2, further comprising calculating the eye openness measures with a tree-based model.

### Claim 4 (depends on 1)

A method for determining user liveness in accordance with claim 1, said calculating confidence scores step comprising:
conducting a series of calculations with the eye openness measures stored in the buffer;
creating a vector from the calculation results; and
calculating the confidence score by processing the vector with a tree-based model.

### Claim 5 (depends on 1)

A method for determining user liveness in accordance with claim 1, said calculating confidence scores step comprising:
determining the location of a temporal window relative to the buffer;
calculating a confidence score based on the eye openness measures included in the temporal window;
shifting the temporal window; and
calculating a second confidence score based on the eye openness measures included in the shifted temporal window.

### Claim 6 (depends on 1)

A method for determining user liveness in accordance with claim 1, further comprising erasing the buffer after said detecting step.

### Claim 7

A system for determining user liveness comprising:
a processor; and
a memory configured to store a buffer of eye openness measures, said processor being programmed to:
calculate eye openness measures for a frame included in captured authentication data, and store the eye openness measures in the buffer;
calculate confidence scores from the eye openness measures stored in the buffer; and
detect an eye blink when a maximum confidence score is greater than a threshold score.

### Claim 8 (depends on 7)

A system for determining user liveness in accordance with claim 7, said processor being further programmed to:
position detection windows included in a table, on a region of interest included in the frame, at coordinates included in the table;
calculate a feature value for each window position; and
calculate the eye openness measures using the feature values.

### Claim 9 (depends on 8)

A system for determining user liveness in accordance with claim 8, said processor being further programmed to calculate the eye openness measures with a tree-based model.

### Claim 10 (depends on 7)

A system for determining user liveness in accordance with claim 7, said processor being further programmed to:
conduct a series of calculations with the eye openness measures stored in the buffer;
create a vector from the calculation results; and
calculate the confidence score by processing the vector with a tree-based model.

### Claim 11 (depends on 7)

A system for determining user liveness in accordance with claim 7, said processor being further programmed to:
determine the location of a temporal window relative to the buffer;
calculate a confidence score based on the eye openness measures included in the temporal window;
shift the temporal window; and
calculate a second confidence score based on the eye openness measures included in the shifted temporal window.

### Claim 12

A method for determining detection windows and window positions to be used for calculating feature values during authentication transactions:
incrementally moving at least one detection window type over a region of interest to occupy different positions within the region of interest;
calculating a feature value for each position;
creating a vector from the feature values;
determining detection window type, size, and position combinations for calculating feature values during authentication transactions; and
determining eye openness measures during authentication transactions with the calculated feature values.

### Claim 13 (depends on 12)

A method for determining a set of detection windows and window positions in accordance with claim 12, said calculating step comprising:
defining different pixel groups with the at least one detection window, the pixel groups being within the region of interest;
calculating an average value for each pixel group; and
calculating the feature value as the difference between the average values.
