Neighborhood Repulsed Metric Learning for Kinship Verification Jiwen - - PowerPoint PPT Presentation

▶
neighborhood repulsed metric learning for kinship
SMART_READER_LITE
LIVE PREVIEW

Neighborhood Repulsed Metric Learning for Kinship Verification Jiwen - - PowerPoint PPT Presentation

Apr 23, 2023 343 likes •654 views

Neighborhood Repulsed Metric Learning for Kinship Verification Jiwen Lu Advanced Digital Sciences Center, Singapore Paper Jiwen Lu, Xiuzhuang Zhou, Yap Peng Tan, Yuanyuan Shang, Jie Zhou, Neighborhood repulsed metric learning for kinship

slide-1
SLIDE 1

Neighborhood Repulsed Metric Learning for Kinship Verification

Jiwen Lu Advanced Digital Sciences Center, Singapore

slide-2
SLIDE 2

Jiwen Lu, Xiuzhuang Zhou, Yap‐Peng Tan, Yuanyuan Shang, Jie Zhou, Neighborhood repulsed metric learning for kinship verification, IEEE Trans. on Pattern Analysis and Machine Intelligence (PAMI), vol. 36, no. 2, pp. 331‐345, 2014. (Conference version appeared at CVPR2012)

Paper

2

slide-3
SLIDE 3

Face Analysis Tasks

Face identification (access control/ surveillance)

3

Face verification (access control/surveillance)

slide-4
SLIDE 4

Face Analysis Tasks

Facial expression recognition (human‐computer interaction)

4

Facial age estimation (visual advertisement/social media)

slide-5
SLIDE 5

Face Analysis Tasks

5

Head pose estimation (human computer interaction) Gender classification (social media analysis)

slide-6
SLIDE 6

Face Analysis Tasks

Facial beauty prediction (multimedia analysis)

6

slide-7
SLIDE 7

Face Analysis Tasks

Kinship verification (social media analysis)

7

Mother‐Daughter (M‐D) Father‐Son (F‐S) Mother‐Son (M‐S) Father‐Daughter (F‐D)

slide-8
SLIDE 8

Related Work

8

Local Features + SVM (Fang2010 [1])

  • 150 image pairs
  • 50%: Caucasians
  • 40%: Asians
  • 7% African Americans
  • 3% others;
  • 40%: F‐S
  • 22%: F‐D
  • 13%: M‐S
  • 26%: M‐D

[1] Ruogu Fang, Kevin D. Tang, Noah Snavely, and Tsuhan Chen, Towards computational models of kinship verification, ICIP, pp. 1577‐1580, 2010. (Best Paper Award)

slide-9
SLIDE 9

Related Work

9

Transfer Learning (Xia2012 [2])

  • 90 groups
  • 3 images in each group

[2] Siyu Xia, Ming Shao, Jiebo Lu, and Yun Fu, Understanding kin relationships in photos, IEEE TMM, vol. 14, no. 8,

  • pp. 1046‐1056, 2012.
slide-10
SLIDE 10

Related Work

10

Salient Feature + SVM (Guo2012 [3])

[3] Guodong Guo and Xiaolong Wang, Kinship measure on salient facial features, IEEE TIM, vol. 61, no. 8, pp. 2322‐ 2325, 2012.

slide-11
SLIDE 11

Datasets

11

  • KinFaceW‐I: 500 kinship image face pairs
slide-12
SLIDE 12

Datasets

12

  • KinFaceW‐II:1000 kinship image face pairs
slide-13
SLIDE 13

Datasets

13

  • Statistics
  • Publicly available: www.kinfacew.com
slide-14
SLIDE 14

Datasets

14

slide-15
SLIDE 15

Mahalanobis Distance

15

Squared Euclidean distance Let The Mahalanobis distance

slide-16
SLIDE 16

Metric Learning

16

Applying Mahalanobis distance to learn a positive semi‐ denite (PSD) matrix Relationship with subspace learning where .

slide-17
SLIDE 17

Representative Metric Learning Algorithms

17

Large Margin Nearest Neighborhood (LMNN)

[Weinberger et al, NIPS2005]

slide-18
SLIDE 18

Representative Metric Learning Algorithms

18

Information‐Theoretic Metric Learning (ITML)

[Davis et al, ICML2007]

where . The optimization function can be re‐formulated as

slide-19
SLIDE 19

Neighborhood Repulsed Metric Learning

19

  • For the verification task, the number of negative pairs is larger

than the number of positive pairs if we know the exact label information of each sample.

  • The importance of different negative pairs is different. Some

negative pairs are very discriminative and some are not so discriminative.

  • It is desirable to identity the most informative negative pairs and

ignore the less informative negative pairs to learn a discriminative metric for verification.

Motivations

slide-20
SLIDE 20

Neighborhood Repulsed Metric Learning

20

[4] Jiwen Lu, Xiuzhuang Zhou, Yap‐Peng Tan, Yuanyuan Shang, Jie Zhou, Neighborhood repulsed metric learning for kinship verification, IEEEE PAMI, vol. 36, no. 2, pp. 331‐345, 2014.

slide-21
SLIDE 21

Multi‐view Neighborhood Repulsed Metric Learning

21

slide-22
SLIDE 22

Multi‐view Neighborhood Repulsed Metric Learning

22

slide-23
SLIDE 23

Experimental Results

23

  • Comparisons with existing metric learning methods

On KinFaceW‐I dataset. On KinFaceW‐II dataset.

slide-24
SLIDE 24

Experimental Results

24

  • Comparisons with multi‐view learning methods
  • Comparisons with human observers

Correct verification accuracy on the KinFaceW‐I dataset. Correct verification accuracy on the KinFaceW‐II dataset.

slide-25
SLIDE 25

Discriminative Multi‐Metric Learning

25

where where

[5] Haibin Yan, Jiwen Lu, Weihong Deng, Xiuzhuang Zhou, Discriminative multimetric learning for kinship verification, IEEE TIFS, vol. 9, no. 7, pp. 1169‐1178, 2014.

slide-26
SLIDE 26

Discriminative Multi‐Metric Learning

26

where Iteration is terminated:

slide-27
SLIDE 27

Discriminative Multi‐Metric Learning

27

The Lagrange function can be constructed as:

slide-28
SLIDE 28

Prototype‐Based Discriminative Feature Learning

28

[5] Haibin Yan, Jiwen Lu, Xiuzhuang Zhou, Prototype‐based discriminative feature learning for kinship verification, IEEE TCYB,2014, accepted.

slide-29
SLIDE 29

29

  • Metric learning is effective for kinship verification.
  • Feature learning for kinship verification?
  • Deep learning for kinship verification?

Summary and Future Work