Грубер Иван Гетерогенное распознавание лиц по эскизам лица

brief description

Грубер Иван Гетерогенное распознавание лиц по эскизам лица
ОГЛАВЛЕНИЕ ДИССЕРТАЦИИ
кандидат наук Грубер Иван
Содержание

Содержание

Реферат

Synopsis

1. Introduction

1.1 Problem definition

1.2 Brief history of Face Recognition

1.3 Motivation and Application

1.4 Goals of Dissertation

1.5 Outline

2. Classification

2.1 Verification

2.2 Identification

2.3 Testing protocols

3. Face Recognition Datasets

3.1 FERET

3.2 XM2VTS

3.3 LFW

3.4 YouTube Faces

3.5 CMU Multi-Pie

3.6 SFC

3.7 CAS-PEAL

3.8 COX Face

3.9 PaSC

3.10 CelebFaces+

3.11 CASIA WebFace

3.12 IJB

3.13 MegaFace

3.14 MS-Celeb-1M

3.15 VGGFace2

3.16 PIPA

3.17 CFP

3.18 CUFS

3.19 CUFSF

3.20 IIIT-D

3.21 Memory Gap Database

3.22 ILSVRC

4. Network Architectures

4.1 Activation functions

4.1.1 Sigmoid

4.1.2 Tanh

4.1.3 ReLU

4.1.4 Leaky ReLU

4.1.5 Parametric ReLU

4.1.6 Maxout

4.1.7 Softmax

4.2 Layers

4.2.1 Fully-connected

4.2.2 Convolutional

4.2.3 Pooling

4.2.4 Normalization

4.2.5 Loss

4.3 Regularization techniques

4.4 Gradient and back-propagation

4.5 Parameter update - optimization methods

4.5.1 SGD

4.5.2 Momentum

4.5.3 Nestorov Momentum

4.5.4 Adagrad

4.5.5 RMSprop

4.5.6 Adam

4.5.7 Nadam

4.6 AlexNet

4.7 VGG

4.8 InceptionNet + NiN

4.9 Highway Networks

4.10 ResNet

4.11 DenseNet

4.12 PyramidalNet

4.13 Squeeze-and-Excitation Networks

4.14 Autoencoders

5. Loss Functions

5.1 Euclidean margin based losses

5.1.1 Softmax loss

5.1.2 Contrastive loss

5.1.3 Triplet loss

5.1.4 Center loss

5.2 Angular and cosine margin based losses

5.2.1 Angular Softmax loss

5.2.2 COCO loss

5.2.3 Arc loss

6. Generative Adversarial Networks

6.1 VAEGAN

6.2 Conditional GAN

6.3 DR-GAN

6.4 FacelD-GAN

6.5 PG-GAN

6.6 Pix2pix

6.7 UNIT/MUNIT

7. Single Image-Based Recognition

7.1 Engineered-based methods

7.1.1 Local feature-based methods

7.1.2 Local appearance-based methods

7.2 Learn-based methods

7.2.1 Statistical methods

7.2.2 AI methods

7.3 3D Face synthesis-based methods

8. Face Recognition from Other Sensory Input

8.1 Video sequence

8.2 Heterogeneous face recognition

8.2.1 Facial Sketches

8.2.2 3D data

8.2.3 Infrared light

9. X-Bridge based heterogeneous face recognition system

9.1 Cross-modal bridge

9.2 Feature Extractor

9.3 Pipeline of the system

9.4 Facial Features Preservation Score

10. Experiments

10.1 Feature extractor comparison

10.1.1 Training data

10.1.2 Comparison of state-of-the-art architectures

10.1.3 Comparison of loss functions

10.1.4 Discussion

10.2 Cross-modal bridge comparison

10.2.1 Training data

10.2.2 Testing data

10.2.3 Quantitative-results testing protocol

10.2.4 PG-GAN

10.2.5 VAEGAN

10.2.6 Pix2pix

10.2.7 UNIT

10.2.8 X-Bridge

10.2.9 Quantitative results comparison

10.2.10 Discussion

11. Conclusion

11.1 Thesis summary

11.2 Dissertation goals

11.2.1 Face recognition methods

11.2.2 Cross-modal bridge comparison

11.2.3 Heterogeneous face recognition system

11.3 Future work

Список рисунков

Список таблиц

Список литературы

Публикации автора по теме диссертации