AxonData commited on
Commit
36d703f
·
verified ·
1 Parent(s): 7edfd5d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +73 -3
README.md CHANGED
@@ -1,3 +1,73 @@
1
- ---
2
- license: cc-by-nc-4.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-nc-4.0
3
+ task_categories:
4
+ - image-classification
5
+ - image-feature-extraction
6
+ task_ids:
7
+ - multi-class-image-classification
8
+ language:
9
+ - en
10
+ tags:
11
+ - face-recognition
12
+ - face-detection
13
+ - face-verification
14
+ - computer-vision
15
+ - biometrics
16
+ - identity-verification
17
+ - kyc
18
+ - portrait
19
+ - demographic-diversity
20
+ - multi-ethnic
21
+ pretty_name: Face Recognition Dataset
22
+ size_categories:
23
+ - 100K<n<1M
24
+ ---
25
+ ## Face Recognition Dataset
26
+ Commercial collection of 100,000+ portrait images with multi-ethnic demographic coverage
27
+
28
+ ## Dataset Summary
29
+ The Axon Labs Face Recognition Dataset is one of the largest commercial face recognition collections available with guaranteed commercial-use licensing, featuring:
30
+ - **100,000+ high-quality face images**
31
+ - **Multi-ethnic demographic coverage** across nationalities and ethnicities
32
+ - **Balanced gender distribution**
33
+ - **Wide age range** from adults to seniors
34
+ - **Varied capture conditions** (devices, lighting, angles)
35
+ - **Commercial license** for production AI systems
36
+ - **GDPR-compliant** with explicit subject consent
37
+
38
+ ## Use Cases
39
+ - **Face Recognition**: train or fine-tune face recognition models on diverse populations
40
+ - **Face Verification**: evaluate 1:1 matching algorithms across demographics
41
+ - **Face Detection**: test detection pipelines on real portrait data
42
+ - **KYC / Identity Verification**: validate models used in remote onboarding
43
+ - **Benchmarking**: compare performance of commercial face recognition systems
44
+ - **Demographic Bias Analysis**: evaluate model fairness across ethnicities, genders, and age groups
45
+
46
+ ## Full Dataset Characteristics
47
+ | Feature | Value |
48
+ |---|---|
49
+ | Number of images | 100,000+ |
50
+ | Number of unique subjects | 100,000+ |
51
+ | Images per subject (avg) | 2-3 |
52
+ | Ethnicities represented | Multiple nationalities and ethnic groups |
53
+ | Gender | Balanced |
54
+ | Age distribution | 14-65+ |
55
+
56
+ ## Why This Dataset
57
+ High-quality face recognition datasets suitable for commercial use are scarce. Most public datasets have restrictive academic-only licenses, lack demographic balance, or contain low-quality images unsuitable for production model training. The Axon Labs Face Recognition Dataset addresses these gaps:
58
+
59
+ - **Scale**: 100,000+ images - one of the largest commercial FR datasets available
60
+ - **Demographic diversity**: multi-ethnic coverage enables fair models across populations
61
+ - **Commercial usability**: licensed for production deployment, not academic-only
62
+ - **Quality over quantity**: every image verified for face visibility and technical quality
63
+ - **Consent-based**: all subjects provided explicit GDPR-compliant consent
64
+ - **Sample available**: evaluate the data before committing to a commercial license
65
+
66
+ ## Related Datasets from Axon Labs
67
+ Axon Labs provides a full catalog of biometric datasets for AI training:
68
+
69
+ - **Selfies and Paired ID Photos Dataset** - 10K+ pairs for selfie-to-ID matching (KYC use case)
70
+ - **Selfie and Video Dataset** - video samples for liveness + face recognition combined
71
+ - **NIST Dataset** - benchmark data for NIST FRVT testing
72
+
73
+ See the full catalog at [axonlab.ai/datasets](https://axonlab.ai/datasets/)