Audio Transcription
Datasets and Services for Training LLM
Image AnnotationAnti-Spoofing Replay Attack Dataset
Replay video attacks from the phone screens to train biometric systems to recognize a real person from a prepared photos and videos
- Liveness detection
- Facial Recognition
- Anti-Spoofing
- iBeta
Our Partners
CASE DESCRIPTION
- Replay attacks on videos from Anti-Spoofing Real with smartphones and monitors
- Each video from Anti-Spoofing Real is filmed on the smartphone camera with zooming
- The dataset includes different smartphones used for attacks, such as iPhone, Xiaomi, Samsung, Vivo
Technical specification
- Each set includes devices used for playing and shooting videos from Anti-Spoofing Real, video resolution, duration and frames per second
- The replay attack duration varies from 10 to 40 seconds, depend on the original video
- Inhouse with various lighting, zooming original video on each attack
- Agreement when providing data for each person
- The dataset meets the requirements of ISO 30107-3 for Presentation Attack
- Type of attack according to the FIDO biometric standard - Displayed videos on mobile devices, Level B
- Data allows you to reduce Attack Presentation Classificaiton Error Rate
-
Anti-Spoofing Real Unique
- Selfie photos and videos from Anti-Spoofing Real, every person in the dataset is unique
- Workers are identified as unique after the verification done by 3 individuals independently
- 38 000 sets with unique people
- Metadata from Anti-Spoofing Real
- Data in 3 days from signing the contract
-
Anti-Spoofing Real
- 98 000+ photos and videos where people turn their heads
- People from 170 countries
- The dataset complies with the requirements of ISO 30107-3 for Bona Fide Presentation
Read more -
Anti-Spoofing Print
- 4 700+ videos with printed A4 photo of a person
- Photos for attacks are taken from Anti-spoofing Real dataset
- The dataset complies with the requirements of ISO 30107-3 for Presentation Attack
- Type of the attack - Printed face image, Level A (FIDO biometric standard)Type of the attack - Printed face image, Level A (FIDO biometric standard)
Read more
APPLICATION AREAS OF THE DATASET
Facial Recognition:
Biometric Identification:
Anti-spoofing:
Once you submit a request, we will contact you to discuss the details and sign the necessary documents. When the documents are signed, you will have the data within 5 days.
We obtain data from both crowdsourcing platforms and our in-house team. The data is further validated by our in-house team.
Upon your request, we can provide part data for any datasets. If you have an interest in specific demographic characteristics, we offer parts from several datasets according to your needs.
We offer additional labeling for an extra fee. For the dataset, we provide special ethnicity classification, lighting classification and other labeling according to your needs.
The price is determined by your requirements. Please submit a request to get a free consultation.
DIDN'T FIND THE NECESSARY INFORMATION?
Why
Training Data
- Quality Assurance:
- Enhanced Data Accuracy
- Consistency in Labels
- Reliable Ground Truth
- Mitigation of Annotation Biases
- Cost and Time Efficiency
- Data Security and Confidentiality:
- GDPR Compliance
- Non-disclosure agreement
- Data Encryption
- Multiple data storage options
- Access Controls and Authentication
- Expert Team:
- 6 years in industry
- 35 top project managers
- 40+ languages
- 100+ countries
- 250k+ assessors
- Flexible and Scalable Solutions:
- 24/7 availability of customer service
- 100% post payment
- $550 minimum check
- Variable Workload
- Customized Solutions