발표자: 홍성은 (SK T-Brain)
발표일: 2018.4.
Real-world face recognition using a single sample per person (SSPP) is a challenging task. The problem is exacerbated if the conditions under which the gallery image and the probe set are captured are completely different. To address these issues from the perspective of domain adaptation, we introduce an SSPP domain adaptation network (SSPP-DAN). In the proposed approach, domain adaptation, feature extraction, and classification are performed jointly using a deep architecture with domain-adversarial training. However, the SSPP characteristic of one training sample per class is insufficient to train the deep architecture. To overcome this shortage, we generate synthetic images with varying poses using a 3D face model. Experimental evaluations using a realistic SSPP dataset show that deep domain adaptation and image synthesis complement each other and dramatically improve accuracy. Experiments on a benchmark dataset using the proposed approach show state-of-the-art performance.
Axa Assurance Maroc - Insurer Innovation Award 2024
Deep Domain Adaptation Network for Face Recognition with Single Sample Per Person
1. SSPP-DAN: Deep Domain Adaptation Network
for Face Recognition with Single Sample Per Person
T-Brain, AI Research Center
Sungeun Hong
2. Contents
Motivation & Problem Definition
Single sample per person (SSPP)
Challenges in real-world face recognition
Proposed method
Domain adaptation
Face synthesis
Experiments
New heterogeneous dataset
LFW for SSPP
An overview of Domain Adaptation
2
4. SSPP face recognition
Face recognition using Single Sample Per Person (SSPP)
Identify or verify identities using only one single gallery image
Related to the recently attracted one-shot learning
4
Sample images of the AR database
ExpressionIllumination
Disguise Multiple
Train (on single gallery)
Test (on probe images)
5. SSPP face recognition
Limitations of most existing SSPP datasets
Lab controlled environment
Shooting environments (e.g., shooting location, camera) are same
5
Lu, Jiwen, Yap-Peng Tan, and Gang Wang. "Discriminative multimanifold analysis for face recognition from a single
training sample per person." IEEE transactions on pattern analysis and machine intelligence 35.1 (2013): 39-51.
Gallery
(train)
Probe
(test)
6. Real-world SSPP Face recognition
6
Registration
(Gallery)
Identification
(Probe)
Gallery
A stable image like
clear frontal mugshot
e.g., ID card or e-passport
Probe
Unstable images including
non-trivial variations
e.g., surveillance camera, web images
Variations:
camera sensor, blur,
noise, pose, illumination
7. Real-world SSPP Face recognition
Challenges
1. Heterogeneity of the shooting environments
Gallery: stable environment
Probe: highly unstable environment
2. Shortage of training samples
Only one training sample per person is available
7
Registration
(Gallery)
Identification
(Probe)
9. Real-world SSPP Face recognition
Challenges
1. Heterogeneity of the shooting environments
Gallery: stable environment
Probe: highly unstable environment
9
Gap
Registration
(Gallery)
Identification
(Probe)
10. Real-world SSPP Face recognition
10
Domain
Adaptation
Registration
(Gallery)
Identification
(Probe)
Challenges
1. Heterogeneity of the shooting environments
Gallery: stable environment
Probe: highly unstable environment
11. Adjust a model from the source domain knowledge to a
different but related target domain
Domain Adaptation (DA)
11
With label
Source domain Target domain
?
Training Test
12. Adjust a model from the source domain knowledge to a
different but related target domain
Unsupervised DA: no label in target domain
Domain Adaptation (DA)
12
Without labelWith label
Source domain Target domain Target domain
?
Training Test
18. Adjust a model to a different target domain starting from
the source domain knowledge
Domain Adaptation (DA)
Model
Source
domain
w/ label
Target
domain
w/o label
18
Model
19. Adjust a model to a different target domain starting from
the source domain knowledge
Domain Adaptation (DA)
Model
Source
domain
w/ label
Target
domain
w/o label
19
Domain
Adaptation
Model
20. Adjust a model to a different target domain starting from
the source domain knowledge
Domain Adaptation (DA)
Model
Source
domain
w/ label
Target
domain
w/o label
Better
result
20
Fine
classification
Domain
Adaptation
Model
21. Domain Adaptation (DA)
Model
Source
domain
w/ label
Target
domain
w/o label
Better
result
21
Fine
classification
Domain
Adaptation
Model
Purpose Train Test
Domain Source Target Target
Image condition Stable Unstable Unstable
Label O X -
22. Domain Adaptation (DA)
22
Basic assumptions of DA
sample distribution of each domain is
related but different (o)
samples are abundant in each domain (x)
Registration
(Gallery)
Identification
(Probe)
Domain
Adaptation
28. Real-world SSPP Face recognition
28
Registration
(Gallery)
Identification
(Probe)
Challenges
1. Heterogeneity of the shooting environments
Gallery: stable environment
Probe: highly unstable environment
2. Shortage of training samples
Only one training sample per person is available
29. SSPP-DAN:
1. Domain adaptation network
From stable face domain (source)
to unstable face domain (target)
2. Face synthesis
Generate virtual samples
29
Domain Adaptation Network for Single Sample Per Person
30. SSPP-DAN
1. Domain adaptation network
with domain-adversarial training
Feature learning
Domain adaptation
Classifier learning
30Ganin et al,"Unsupervised domain adaptation by backpropagation." ICML. 2015.
Jointly
Source
or
target?
Class
label?
31. SSPP-DAN
1. Domain adaptation network
with domain-adversarial training
Feature learning
Domain adaptation
Classifier learning
31Ganin et al,"Unsupervised domain adaptation by backpropagation." ICML. 2015.
Jointly
Source
or
target?
Class
label?
32. SSPP-DAN
1. Domain adaptation network
with domain-adversarial training
Feature learning
Domain adaptation
Classifier learning
32Ganin et al,"Unsupervised domain adaptation by backpropagation." ICML. 2015.
Jointly
Source
or
target?
Class
label?
33. SSPP-DAN
33
1) Landmark detection
• Supervised descent method
2) 2D → 3D mapping
3) Pose estimation
4) Image synthesis
• (yaw: -80°~+80°, pitch: -10°~40°)
1. Domain adaptation network
From stable face domain (source)
to unstable face domain (target)
2. Face synthesis
Generate virtual samples
41. Labeled Faces in the Wild (LFW)
41
Source Target
Dataset summary
Rearranged LFW-A for SSPP face recognition
Gallery: 50 images for 50 people
Generic set: 108 subjects
42. Labeled Faces in the Wild (LFW)
42
Source Target
LFW for SSPP protocol
Gallery: 50 images for 50 people
Generic set: 108 subjects
Deep
learning
43. Summary
43
Challenges
1. Heterogeneity of the shooting environments
2. Shortage of training samples
Solution
Deep neural network with domain adversarial training
Facial image synthesis
Complement
each other
45. Domain Adaptation
Baseline and variants
Domain-invariant feature learning
Pixel DA using GAN
Application
Classification
Segmentation
Detection
Etc.
45