Fake Fingerprint Defensive Power Assessment_Union Community_ENG

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

Union Community Co., Ltd.

Fingerprint Security Token

Fake Fingerprint Defensive Power Assessment Report

2010. 05.

YONSEI University

Biometrics Engineering Research Center

Designated by the Korea Science Engineering Foundation

http://berc.yonsei.ac.kr

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

Contents

Chapter 1. Introduction (1)

Section 1. Fake Fingerprint Defensive Power Assessment Scenario

Chapter 2. Defensive Power Performance Assessment Scale and Contents

Section 1. Defensive Power Assessment Based Information (4)

Chapter 3. Defensive Power Assessment Result and Analysis

Section1. Defesnive Power Assessment Result (9)

Section2. Defensive Power Assessment Result Analysis (19)

Chapter 4. Fake Fingerprint Manufacturing Process & Quality Assessment Section 1. Fake Fingerprint Quality Assessment (22)

References (23)

2

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

Chapter 1. Introduction

Section 1. Fake Fingerprint Defensive Power Assessment Scenario

According to the survey of various types of fingerprint security token attack methods using fake fingerprints, the system needs to be assessed by distinguishing the following 4 scenarios (Table 1) in order to assess the defensive power of fingerprint security token against fake fingerprint. [1, 2]

Scenario Enrollment Authentication

Scenario A Real fingerprint Real fingerprint

Scenario B Real fingerprint Fake fingerprint

Scenario C Fake fingerprint Real fingerprint

Scenario D Fake fingerprint Fake fingerprint Table 1. Four Types of Fingerprint Security Token Defensive Power Assessment

⊙ Scenario A: As the normal type of fingerprint authentication, a real user registers and authenticates fingerprint. This assessment type is used as reference data to provide reliability for defensive power assessment.

⊙ Scenario B: As the most general and important scenario of fake fingerprint attack, this scenario represents a case where a system attacker intrudes the system by creating a fake fingerprint of a real user.

⊙ Scenario C: This scenario represents a case where a system user registers his/her own fake fingerprint using a representative to hide identity and uses his/her own real fingerprint during authentication. Note that a fake fingerprint is used during the enrollment process.

⊙ Scenario D: A real user uses this method to hide his/her id entity. Note that a fake fingerprint is created and used during both enrollment and authentication steps.

Therefore, to assess the defensive power of the fingerprint security token against fake fingerprint, the system’s enrollment and authentication steps need to be considered. In addition, a defensive power assessment method for real fingerprint and fake fingerprint needs to be provided in each step.

3

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

Chapter 2. Defensive Power Performance Assessment Scale

Section 1. Defensive Power Assessment Scale and Contents

1. Enrollment Process Assessment Scale

1.1 Assessment Scale during Real Fingerprint Enrollment

⊙ Non‐Response Rate (NRR)

‐ Rate of real fingerprint inputs not entered due to the unique characteristics of sensor

‐ NRR = (The number of real fingerprint inputs not entered in the sensor / The total number of real fingerprint input trials)

‐ Example) Case where the system does not respond after a real fingerprint input is made, or Case where the system does not respond within the allowed input time after a real fingerprint input is made (Time out)

⊙ Submission Rejection Rate (SRR)

‐ Rate of re‐input requests due to the poor quality of a real fingerprint or inaccurate fingerprint input location after the sensor receives fingerprint input

‐ SRR = (The number of re‐input requests during real fingerprint input / The total number of real fingerprint input trials)

‐ Example) Re‐input request by the system due to poor image quality of entered real fingerprint, or Re‐input request by the system due to inad equate rubbing input speed of a real fingerprint (Rubbing characteristics of a sweep sensor)

4

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

1.2 Assessment Scale during Fake Fingerprint Enrollment

⊙ Non‐Response Rate (NRR)

‐ Rate of fake fingerprint inputs not entered due to the unique characteristics of sensor

‐ NRR = (The number of fake fingerprint inputs not entered in the sensor / The total number of fake fingerprint input trials)

‐ Example)

Case where the system does not respond during fake fingerprint input

Case where the system generates input time exceeding (time out) error during fake fingerprint input

⊙ Submission Rejection Rate (SRR)

‐ Rate of re‐input requests due to the poor quality of fake fingerprint or inaccurate input location after the sensor receives fingerprint input

‐ SRR = (The number of re‐input requests during fake fingerprint input / The total number of fake fingerprint input trials)

‐ Example)

Re‐input request by the system due to poor quality of entered fake fingerprint,

Re‐input request due to inadequate input speed of fake fingerprint

⊙ Spoof Detection Rate (SDR)

‐ Rate of the system deciding fake fingerprints correctly as fake fingerprint

‐ SDR = (Rate of the system deciding fake fingerprints correctly as fake fingerprint / The total number of fake fingerprint input trials)

‐ Example)

Case where the system detects a fake fingerprint during fake fingerprint input

⊙ Spoof Enrollment Count (SEC)

‐ The number of successful enrollments by fake fingerprint

‐ SEC = (The total number of successful enrollments by fake fingerprints)

‐ Example)

Case where fake fingerprint enrollment eventually becomes successful when the system tries to enroll once by accepting a large number of fingerprint images

⊙ Spoof Acceptance Rate (SAR)

‐ Rate that accepts fake fingerprints incorrectly as real fingerprints

‐ SAR = (The number of times that accept fake fingerprint input incorrectly as real fingerprint / The total number of fake fingerprint input trials)

5

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

‐ Example)

Case of successful input during fake fingerprint enrollment

(*Therefore, Non‐Response Rate (NRR) + Submission Rejection Rate (SRR) + Spoof Detection Rate (SDR) + Spoof Acceptance Rate (SAR) = 1. Spoof Enrollment Rate (SEC) is the number of final enrollments through a large number of successful fake fingerprint inputs. It is affected by the number of inputs required during the enrollment of the assessment system.)

2. Authentication Process Assessment Scale

2.1 Assessment Scale during Real Fingerprint Authentication

⊙ Non‐Response Rate (NRR)

‐ Rate of real fingerprints not entered in the sensor due to the unique characteristics of sensor ‐ NRR = (The number of real fingerprint inputs not entered in the sensor / The total number of real fingerprint input trials)

‐ Example)

Case where the system does not respond after real fingerprint input is made, Case where the system issues the input time exceeding (time out) error during real

fingerprint input

⊙ Submission Rejection Rate (SRR)

‐ Rate of re‐input requests due to the poor quality of real fingerprint or inaccurate input location after the sensor receives fingerprint input

‐ SRR = (The number of re‐input requests during real fingerprint input / The total number of real fingerprint re‐input trials)

‐ Example)

Re‐input request by the system due to poor quality of entered real fingerprint

Re‐input request by the system due to inadequate input speed of real fingerprint

2.2 Assessment Scale during Fake Fingerprint Authentication

⊙ Non‐Response Rate (NRR)

‐ Rate of fake fingerprints not entered in the sensor due to the unique characteristics of sensor ‐ NRR = (The number of fake fingerprint inputs not entered in the sensor / The total number of fake fingerprint input trials)

‐ Example)

Case where the system does not respond after fake fingerprint input is made

6

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

Case where the system issues the input time exceeding (time out) error during fake fingerprint input

⊙ Submission Rejection Rate (SRR)

‐ Rate of re‐input requests due to the poor quality of fake fingerprint or inaccurate input location after the sensor receives fingerprint input

‐ SRR = (The number of re‐input requests during fake fingerprint input / The total number of fake fingerprint re‐input trials)

‐ Example)

Re‐input request by the system due to poor quality of entered fake fingerprint

Re‐input request by the system due to inadequate input speed of fake fingerprint

⊙ Spoof Detection Rate (SDR)

‐ Rate of the system deciding fake fingerprints correctly as fake fingerprint

‐ SDR = (Rate of the system deciding fake fingerprints correctly as fake fingerprint / The total number of fake fingerprint input trials)

‐ Example)

Case where the system detects a fake fingerprint during fake fingerprint input

⊙ Spoof Acceptance Rate (SAR)

‐ Rate that accepts fake fingerprints incorrectly as real fingerprints

‐ SAR = (The number of times that accept fake fingerprint input incorrectly as real fingerprint / The total number of fake fingerprint input trials)

‐ Example)

Case of successful input during fake fingerprint enrollment

Therefore, using the above defensive power performance assessment scal es, the defensive power performance of the fingerprint security token against each fake fingerprint can be defined as in the following expression.

Defensive Power Performance = 1 – Spoof Acceptance Rate (SAR) / (Non‐Response Rate (NRR) + Submission Rejection Rate (SRR) + Spoof Detection Rate (SDR) + Spoof Acceptance Rate (SAR))

This value is obtained by dividing ‘The number of times that accept fake fingerprints incorrectly as real fingerprint’ with ‘The total number of input trials’ and then subtracting the result from 1.

It ranges between 0 and 1. Therefore, a high defensive power performance value obtained through the above formula represents that the fingerprint security token has strong defensive

7

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

power against potential fake fingerprints (paper, silicone, and others). On the contrary, a low defensive power performance value represents that the fingerprint security token has poor defensive power against potential fake fingerprints.

8

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

Chapter 3. Defensive Power Assessment Result and Analysis

In this Chapter 3, it indicates the defensive power assessment result of Fingerprint Security Token (Bio‐seal) manufactured by Union Community Co., Ltd.

Section 1. Defensive Power Assessment Result

1. Scenario A (Real Finger Enrollment and Real Finger Authentication)

1.1 Enrollment Step (Real Fingerprint)

Assessment Scal e

Participant

Non‐Response Rate (%) Submission Rejection Rate (%) The Total Number

of Trials

1 0% (0/40) 0% (0/40) 40

2 0% (0/40) 0% (0/40) 40

3 0% (0/40) 0% (0/40) 40

4 0% (0/40) 0% (0/40) 40

5 5% (2/40) 0% (0/40) 40

Sum 1% (2/200) 0% (0/200) 200

1.2 Authentication Step (Real Fingerprint)

Assessment Scal e

Participant

Non‐Response Rate (%) Submission Rejection Rate (%) The Total Number

of Trials

1 0% (0/40) 0% (0/40) 40

2 0% (0/40) 0% (0/40) 40

3 0% (0/40) 0% (0/40) 40

4 0% (0/40) 0% (0/40) 40

5 0% (0/40) 0% (0/40) 40

Sum 0% (0/200) 0% (0/200) 200

9

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

10

2. Scenario B (Real Fingerprint Enrollment and Fake Fingerprint Authentication)

2.1 Enrollment Step (Real Fingerprint)

Assessment Scal e

Participant Non‐Response Rate (%) Submission Rejection Rate (%)

The Total Number

of Trials 1 0% (0/40) 0% (0/40) 40 2 0% (0/40) 0% (0/40) 40 3 0% (0/40) 0% (0/40) 40 4 0% (0/40) 0% (0/40) 40 5 0% (0/40) 0% (0/40) 40 Sum

0% (0/200) 0% (0/200)

200

2.2 Authentication Step (Fake Fingerprint) ‐ Paper Fake Fingerprint

Assessment Scal e

Fake Type

Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%) The Total Number

of Trials 1‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 1

1‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 2‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 2

2‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 3‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 3

3‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 4‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 4

4‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 5‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 Paper

5 5‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 Sum

100% (200/200)

0% (0/200)

0% (0/200)

0% (0/200)

200

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

11

‐ OHP Film Fake Fingerprint

Assessment Scal e

Fake Type

Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%) The Total Number

of Trials 1‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 1

1‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 2‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 2

2‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 3‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 3

3‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 4‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 4

4‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 5‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 OHP Film

5 5‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 Sum

100% (200/200)

0% (0/200)

0% (0/200)

0% (0/200)

200

‐ Rubber Fake Fingerprint

Assessment Scal e

Fake Type

Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%) The Total Number

of Trials 1‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 1

1‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 2‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 2

2‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 3‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 3

3‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 4‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 4

4‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 5‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 Rubber

5 5‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 Sum

100% (200/200)

0% (0/200)

0% (0/200)

0% (0/200)

200

‐ Silicone Fake Fingerprint

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

12

Assessment Scal e

Fake Type

Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%) The Total Number

of Trials 1‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 1

1‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 2‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 2

2‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 3‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 3

3‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 4‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 4

4‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 5‐1 90% (18/20) 10% (2/20) 0% (0/20) 0% (0/20) 20 Silicone 5 5‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 Sum

99% (198/200)

1% (2/200)

0% (0/200)

0% (0/200)

200

‐ Gelatin Fake Fingerprint

Assessment Scal e

Fake Type

Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%) The Total Number

of Trials 1‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 1

1‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 2‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 2

2‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 3‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 3

3‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 4‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 4

4‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 5‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 Gelatin

5 5‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 20 Sum

100% (200/200)

0% (0/200)

0% (0/200)

0% (0/200)

200

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

13

‐ Prosthetic Hand Fake Fingerprint

Assessment Scal e

Fake Type Participant Sampl e

NRR (%) SRR (%) SDR (%)SAR (%) The Total Number

of Trials 1 1‐1 100% (40/40) 0% (0/40) 0% (0/40) 0% (0/40) 40 2

2‐1 100% (40/40) 0% (0/40) 0% (0/40) 0% (0/40) 40 3 3‐1 100% (40/40)

0% (0/40) 0% (0/40) 0% (0/40) 40 4 N/A ‐ ‐ ‐ ‐ ‐ Prosthetic Hand

5 N/A

‐ ‐ ‐ ‐ ‐ Sum

100% (120/120)

0% (0/120)

0% (0/120)

0% (0/120)

120

3. Scenario C (Fake Fingerprint Enrollment and Real Fingerprint Authentication) 3.1 Enrollment Step (Fake Fingerprint) ‐ Paper Fake Fingerprint

Assessment Scal e

Fake Type

Participant Sampl e

NRR (%)

SRR (%)

SDR (%) SAR (%) # of Fake

Enrollments The Total Number

of Trials

1‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 1

1‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 2‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 2

2‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 3‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 3

3‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 4‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 4

4‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 5‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 Paper 5 5‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 Sum

100% (200/200) 0% (200/200)

0% (200/200)

0% (200/200)

200

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

‐ OHP Film Fake Fingerprint

Assessment Scal e

Fake Type Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%)

# of Fake

Enrollments

The Total

Number of

Trials 1‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 1

1‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 2‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 2

2‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 3‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3

3‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

4‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

4

4‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 5‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

OHP Film

5

5‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Sum

100%

(200/200)

0%

(200/200)

0%

(200/200)

0%

(200/200)0 200

‐ Rubber Fake Fingerprint

Assessment Scal e

Fake Type Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%)

# of Fake

Enrollments

The Total

Number

of Trials 1‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 1

1‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

2‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

2

2‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3

3‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 4‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 4

4‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 5‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Rubber

5

5‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Sum

100%

(200/200)

0%

(200/200)

0%

(200/200)

0%

(200/200)

0 200

14

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

‐ Silicone Fake Fingerprint

Assessment Scal e

Fake Type Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%)

# of Fake

Enrollments

The Total

Number

of Trials 1‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 1

1‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 2‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 2

2‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 3‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3

3‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

4‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

4

4‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 5‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Silicone

5

5‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Sum

100%

(200/200)

0%

(200/200)

0%

(200/200)

0%

(200/200)0 200

‐ Gelatin Fake Fingerprint

Assessment Scal e

Fake Type Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%)

# of Fake

Enrollments

The Total

Number

of Trials 1‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 1

1‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

2‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

2

2‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3

3‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 4‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 4

4‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 5‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Gelatin

5

5‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Sum

100%

(200/200)

0%

(200/200)

0%

(200/200)

0%

(200/200)

0 200

15

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

16

‐ Prosthetic Hand Fake Fingerprint

Assessment Scal e

Fake Type Participant Sampl e

NRR (%) SRR (%) SDR (%)SAR (%) # of Fake Enrollments

The Total Number of Trials 1 1‐1 100% (40/40) 0% (0/40) 0% (0/40) 0% (0/40) 0 40 2

2‐1 100% (40/40) 0% (0/40) 0% (0/40) 0% (0/40) 0 40 3 3‐1 100% (40/40)

0% (0/40) 0% (0/40) 0% (0/40) 0 40 4 N/A ‐ ‐ ‐ ‐ ‐ ‐ Prosthetic Hand

5 N/A

‐ ‐ ‐ ‐ ‐ ‐ Sum

100% (120/120)

0% (0/120)

0% (0/120)

0% (0/120)

120

3.2 Authentication Step (Real Fingerprint)

‐ The authentication assessment was not able to be performed as all paper, OHP Film, Rubber, Silicone, Gelatin and Prosthetic Hand fake fingerprints were not enrolled.

4. Scenario D (Fake Fingerprint Enrollment, Fake Fingerprint Authentication) 4.1 Enrollment Step (Fake Fingerprint) ‐ Paper Fake Fingerprint

Assessment Scal e

Fake Type

Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%) # of Fake

Enrollments The Total Number

of Trials

1‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 1

1‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 2‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 2

2‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 3‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 3

3‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 4‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 4

4‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 5‐1 100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 Paper 5 5‐2

100% (20/20) 0% (0/20) 0% (0/20) 0% (0/20) 0 20 Sum

100% (200/200)

0% (200/200)

0% (200/200)

0% (200/200)

200

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

‐ OHP Film Fake Fingerprint

Assessment Scal e

Fake Type Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%)

# of Fake

Enrollments

The Total

Number

of Trials 1‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 1

1‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 2‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 2

2‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 3‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3

3‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

4‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

4

4‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 5‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

OHP Film

5

5‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Sum

100%

(200/200)

0%

(200/200)

0%

(200/200)

0%

(200/200)0 200

‐ Rubber Fake Fingerprint

Assessment Scal e

Fake Type Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%)

# of Fake

Enrollments

The Total

Number

of Trials 1‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 1

1‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

2‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

2

2‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3

3‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 4‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 4

4‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 5‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Rubber

5

5‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Sum

100%

(200/200)

0%

(200/200)

0%

(200/200)

0%

(200/200)

0 200

17

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

‐ Silicone Fake Fingerprint

Assessment Scal e

Fake Type Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%)

# of Fake

Enrollments

The Total

Number

of Trials 1‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 1

1‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 2‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 2

2‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 3‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3

3‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

4‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

4

4‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 5‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Silicone

5

5‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Sum

100%

(200/200)

0%

(200/200)

0%

(200/200)

0%

(200/200) 0 200

‐ Gelatin Fake Fingerprint

Assessment Scal e

Fake Type Participant Sampl e

NRR (%) SRR (%) SDR (%) SAR (%)

# of Fake

Enrollments

The Total

Number

of Trials 1‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 1

1‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

2‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

2

2‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20)

0 20

3

3‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 4‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 4

4‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20 5‐1

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Gelatin

5

5‐2

100%

(20/20)

0%

(0/20)

0%

(0/20)

0%

(0/20) 0 20

Sum

100%

(200/200)

0%

(200/200)

0%

(200/200)

0%

(200/200)

0 200

18

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

19

‐ Prosthetic Hand Fake Fingerprint

Assessment Scal e

Fake Type Participant Sampl e

NRR (%) SRR (%) SDR (%)SAR (%) # of Fake Enrollments

The Total Number of Trials 1 1‐1 100% (40/40) 0% (0/40) 0% (0/40) 0% (0/40) 0 40 2

2‐1 100% (40/40) 0% (0/40) 0% (0/40) 0% (0/40) 0 40 3 3‐1 100% (40/40)

0% (0/40) 0% (0/40) 0% (0/40) 0 40 4 N/A ‐ ‐ ‐ ‐ ‐ ‐ Prosthetic Hand

5 N/A

‐ ‐ ‐ ‐ ‐ ‐ Sum

100% (120/120)

0% (0/120)

0% (0/120)

0% (0/120)

120

4.2 Authentication Step (Fake Fingerprint)

‐ Since paper, OHP film, rubber, silicone, gelatin, and prosthetic hand are not enrolled, therefore, real fingerprint authentication assessment is not possible.

Section 2. Defensive Power Assessment Result Analysis

◎ Scenario‐by‐scenario Analysis

⊙ Scenario A (Real Fingerprint Enrollment, Real Fingerprint Authentication)

‐ The submission rejection rate in enrollment and authentication steps is 0%. ‐ The non‐response rate in enrollment step is 1%.

‐ Analysis: The case of the system not responding to a real fingerprint occurred once, but it

is extremely rare. Therefore, it is thought to be acceptable to enroll and use a real person’s fingerprint instead of a fake fingerprint.

⊙ Scenario B (Real Fingerprint Enrollment, Fake Fingerprint Authentication)

‐ Enrollment Step

‐ The non‐response rate and submission rejection rate of a real fingerprint are 0%.

‐ Analysis: In case of a real person’s fingerprints, the input of all fingerprints was allowed to

enable enrollment. ‐ Authentication Step

‐ The spoof acceptance rate by fake fingerprints of paper, OHP film, rubber, gelatin, and prosthetic hand is 0%. When enrolling real fingerprints and authenticating with the above fake fingerprints, authentication in all occasions and the non‐response rate was 100%. Therefore, it can be interpreted that detecting the above fake fingerprints in terms of hardware was attempted.

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

‐ The non‐response rate and submission rejection rate against silicone fake fingerprints were 99% and 1%, respectively. Enrolling with real fingerprints and authenticating with silicone fingerprints failed in all occasions.

‐ Therefore, defensive power performance for scenario B becomes the value of 1 against all fake fingerprints.

Fake Type NRR (%) SRR (%) SDR (%) SAR (%)

Defensive

Power Performance

Paper 100% 0% 0% 0% 1

OHP Film 100% 0% 0% 0% 1

Rubber 100% 0% 0% 0% 1

Silicone 99% 1% 0% 0% 1

Gelatin 100% 0% 0% 0% 1

Prosthetic

Hand

100% 0% 0% 0% 1

⊙ Scenario C (Fake Fingerprint Enrollment, Real Fingerprint Authentication)

‐ Enrollment for fingerprints of paper, OHP film, silicone, gelatin, rubber, and prosthetic hand failed. Therefore, authentication assessment foll owing it is not possible.

⊙ Scenario D (Fake Fingerprint Enrollment, Fake Fingerprint Authentication)

‐ Enrollment for fingerprints of paper, OHP film, silicone, gelatin, rubber, and prosthetic hand failed. Therefore, authentication assessment foll owing it is not possible.

‐ Therefore, in defensive power performance enrollment and authentication steps for scenarios C and D, the defensive power performance has the value of 1 against all fake fingerprints.

Fake Type NRR (%) SRR (%) SDR (%) SAR (%)

Defensive

Power Performance

Paper 100% 0% 0% 0% 1 OHP Film 100% 0% 0% 0% 1 Rubber 100% 0% 0% 0% 1 Silicone 100% 0% 0% 0% 1 Gelatin 100% 0% 0% 0% 1 Prosthetic

Hand

100% 0% 0% 0% 1

20

Fake Fingerprint Defensive Power Assessment_Union Community_ENG

◎ Review and Comment

‐ In case of scenario B (real fingerprint enrollment/fake fingerprint authentication), the tested Union Community’s fingerprint security token has the spoof acceptance rate of 0% and defensive power performance value of 1 against fake fingerprints of paper, OHP film, rubber, silicone, gelatin, and prosthetic hand. Therefore, for this fingerprint security token, the authentication with general fake fingerprints as well as fake fingerprints used in this test is determined to be very difficult.

‐ In case of scenario C (fake fingerprint enrollment/real fingerprint authentication) and scenario D (fake fingerprint enrollment/fake fingerprint authentication), the enrollment of all fake fingerprints was not allowed. As a result, authentication assessment could not be performed. Also, defensive power performance was 1.

‐ Therefore, the Union Community’s fingerprint security token was found to have a very strong defensive power against fake fingerprints in all scenarios.

21

本文来源：https://www.bwwdw.com/article/77m4.html