AxonData commited on
Commit
aafb41e
·
verified ·
1 Parent(s): f3b2a20

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -23,9 +23,9 @@ tags:
23
  The **iBeta Level 1 Certification Dataset** focuses on **paper mask attacks** tested during **iBeta Level 1** **Presentation Attack Detection (PAD)**. This dataset includes multiple variations of paper mask attacks for training AI models to distinguish between real and spoofed facial data, and it is tailored to meet the requirements for iBeta certifications.
24
 
25
  ### Key Features
26
- - **40+ Participants**: Engaged in the dataset creation, with a balanced representation of **Caucasian, Black, and Asian** ethnicities.
27
  - **Video Capture**: Videos are captured on **iOS and Android phones**, featuring **multiple frames** and **approximately 10 seconds** of video per attack.
28
- - **18,000+ Paper Mask Attacks**: Including a variety of attack types such as print and cutout paper masks, cylinder-based attacks to create a volume effect, and 3D masks with volume-based elements (e.g., nose).
29
  - **Active Liveness Testing**: Includes a **zoom-in and zoom-out phase** to simulate **active liveness detection**.
30
  - **Variation in Attacks**:
31
  - Real-life selfies and videos from participants.
 
23
  The **iBeta Level 1 Certification Dataset** focuses on **paper mask attacks** tested during **iBeta Level 1** **Presentation Attack Detection (PAD)**. This dataset includes multiple variations of paper mask attacks for training AI models to distinguish between real and spoofed facial data, and it is tailored to meet the requirements for iBeta certifications.
24
 
25
  ### Key Features
26
+ - **80+ Participants**: Engaged in the dataset creation, with a balanced representation of **Caucasian, Black, and Asian** ethnicities.
27
  - **Video Capture**: Videos are captured on **iOS and Android phones**, featuring **multiple frames** and **approximately 10 seconds** of video per attack.
28
+ - **22,000+ Paper Mask Attacks**: Including a variety of attack types such as print and cutout paper masks, cylinder-based attacks to create a volume effect, and 3D masks with volume-based elements (e.g., nose).
29
  - **Active Liveness Testing**: Includes a **zoom-in and zoom-out phase** to simulate **active liveness detection**.
30
  - **Variation in Attacks**:
31
  - Real-life selfies and videos from participants.