以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary...

60
以以以以以以以以以以以以以以以以以以以以 Face Recognition with Local Binary Patterns and Partial Matching Presenter: 以以以 Advisor: 以以以 以以 1

description

以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns and Partial Matching. Presenter : 施佩汝 Advisor : 歐陽明 教授. Outlines. Motivation Implementation Result Conclusion. Motivation. Motivation. Publication. - PowerPoint PPT Presentation

Transcript of 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary...

Page 1: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

以區域二元圖樣與部分比對為基礎之人臉辨識Face Recognition with Local Binary

Patterns and Partial Matching

Presenter: 施佩汝Advisor: 歐陽明 教授

1

Page 2: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Outlines• Motivation• Implementation• Result• Conclusion

2

Page 3: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

MOTIVATION

3

Page 4: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Motivation

4

Page 5: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Publication• Che-Hua Yeh, Pei-Ruu Shih, Kuan-Ting Liu, Yin-Tzu

Lin, Huang-Ming Chang, Ming Ouhyoung. A Comparison of Three Methods of Face Recognition for Home Photos. ACM Siggraph, poster, 2009.

5

Page 6: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Problem Statement

6

Page 7: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Main Contribution• Improve Local Binary Patterns by using Partial

Matching Metric

• Better Performance in Home Photos

7

Page 8: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

IMPLEMENTATION

8

Page 9: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

System Overview

Pre-Processing

Build Descriptor

Images Descriptors

ClusteringCalculate

LBPBuild

Descriptor

9

Page 10: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

System Overview

Pre-Processing

Build Descriptor

Images Descriptors

ClusteringCalculate

LBPBuild

Descriptor

10

Page 11: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Pre-Processing

11

Page 12: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

System Overview

Pre-Processing

Build Descriptor

Images Descriptors

ClusteringCalculate

LBPBuild

Descriptor

12

Page 13: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

System Overview

Pre-Processing

Build Descriptor

Images Descriptors

ClusteringCalculate

LBPBuild

Descriptor

13

Page 14: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Local Binary Patterns [PAMI2006]

• An operator to encode the relationship of a pixel and its neighbors. Z1 Z2 Z3

Z8 Z0 Z4

Z7 Z6 Z5

58 99 21

54 54 86

67 12 13

1

otherwise ,00 if ,1

28

10

xxs

ZZsLBP p

pp

14

Page 15: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Local Binary Patterns [PAMI2006]

• An operator to encode the relationship of a pixel and its neighbors.

58 99 21

54 54 86

67 12 13

1 1

otherwise ,00 if ,1

28

10

xxs

ZZsLBP p

pp

Z1 Z2 Z3

Z8 Z0 Z4

Z7 Z6 Z5

15

Page 16: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Local Binary Patterns [PAMI2006]

• An operator to encode the relationship of a pixel and its neighbors.

58 99 21

54 54 86

67 12 13

1 1 0

otherwise ,00 if ,1

28

10

xxs

ZZsLBP p

pp

Z1 Z2 Z3

Z8 Z0 Z4

Z7 Z6 Z5

16

Page 17: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Local Binary Patterns [PAMI2006]

• An operator to encode the relationship of a pixel and its neighbors.

58 99 21

54 54 86

67 12 13

1 1 0

1

otherwise ,00 if ,1

28

10

xxs

ZZsLBP p

pp

Z1 Z2 Z3

Z8 Z0 Z4

Z7 Z6 Z5

17

Page 18: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Local Binary Patterns [PAMI2006]

• An operator to encode the relationship of a pixel and its neighbors.

58 99 21

54 54 86

67 12 13

1 1 0

1

0

otherwise ,00 if ,1

28

10

xxs

ZZsLBP p

pp

Z1 Z2 Z3

Z8 Z0 Z4

Z7 Z6 Z5

18

Page 19: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Local Binary Patterns [PAMI2006]

• An operator to encode the relationship of a pixel and its neighbors.

58 99 21

54 54 86

67 12 13

1 1 0

1

0 0

otherwise ,00 if ,1

28

10

xxs

ZZsLBP p

pp

Z1 Z2 Z3

Z8 Z0 Z4

Z7 Z6 Z5

19

Page 20: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Local Binary Patterns [PAMI2006]

• An operator to encode the relationship of a pixel and its neighbors.

58 99 21

54 54 86

67 12 13

1 1 0

1

1 0 0

otherwise ,00 if ,1

28

10

xxs

ZZsLBP p

pp

Z1 Z2 Z3

Z8 Z0 Z4

Z7 Z6 Z5

20

Page 21: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Local Binary Patterns [PAMI2006]

• An operator to encode the relationship of a pixel and its neighbors.

58 99 21

54 54 86

67 12 13

1 1 0

1 1

1 0 0

otherwise ,00 if ,1

28

10

xxs

ZZsLBP p

pp

Z1 Z2 Z3

Z8 Z0 Z4

Z7 Z6 Z5

21

Page 22: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Local Binary Patterns [PAMI2006]

• An operator to encode the relationship of a pixel and its neighbors.

58 99 21

54 54 86

67 12 13

1 1 0

1 1

1 0 0

otherwise ,00 if ,1

28

10

xxs

ZZsLBP p

pp

Z1 Z2 Z3

Z8 Z0 Z4

Z7 Z6 Z5

LBP = 1101001122

Page 23: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

System Overview

Prepared-Works

Build Descriptor

Images Descriptors

ClusteringCalculate

LBPBuild

Descriptor

23

Page 24: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Facial Image Descriptor• They use Spatially Enhanced Histogram in

original Local Binary Pattern. [PAMI2006]

24

Page 25: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Local Patches• We sample a patch for every s pixels.• There are S patches for one image.

mm

s s

25

Page 26: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Spatial Block [CVPR2007]

• We use three concentric circles to describe a patch.

26

Page 27: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Descriptor• Build a descriptor for one face.

27

KKKK

K

K

fff

ffffff

21

22221

11211

Page 28: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

System Overview

Pre-Processing

Build Descriptor

Images Descriptors

ClusteringCalculate

LBPBuild

Descriptor

28

Page 29: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

System Overview

Pre-Processing

Build Descriptor

Images Descriptors

Complete-Linkage Clustering

Calculate LBP

Build Descriptor

Compute all the similarities Clustering

29

Page 30: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

System Overview

Pre-Processing

Build Descriptor

Images Descriptors

Complete-Linkage Clustering

Calculate LBP

Build Descriptor

Compute all the similarities Clustering

30

Page 31: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Similarity• They use the weighted Chi-Square distance in

original Local Binary Pattern. [PAMI2006]

j ijij

ijij

iiw MS

MSwMS

22 ,

31

Page 32: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Partial Matching [ICCV2009]

• Step1:– Compute the similarity of each patch from one

image with the nearby patches in another image.

,

Image 1: I(1) Image 2: I(2)

)2()1(||;|:|,

)1( ,min klijrslsjrsksilkij ffsimfd

32

Page 33: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Partial Matching [ICCV2009]

• Step2:– Sort the similarities of all patches.

– dαS is the similarity of I(1) to I(2). K

jiijsS fdSortdddd1,

)1(21 ,...,,...,,

SdIId )( )2()1(

33

Page 34: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Partial Matching [ICCV2009]

• Step3:– Calculate the similarity of I(2) to I(1)

)( )1()2( IId

34

Page 35: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Partial Matching [ICCV2009]

• Step4:– Use the maximum of two similarity

)(),(max),( )1()2()2()1()2()1( IIdIIdIID

35

Page 36: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

System Overview

Pre-Processing

Build Descriptor

Images Descriptors

Complete-Linkage Clustering

Calculate LBP

Build Descriptor

Compute all the similarities Clustering

36

Page 37: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Hierarchical Clustering• Build a tree-based hierarchical taxonomy

(dendrogram) from a set of documents.

Material Selected from Introduction to Information Retrieval, Christopher D. Manning, Prabhakar Raghavan and Hinrich Schütze, 2008.

37

Page 38: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Hierarchical Clustering• Clustering obtained by cutting the dendrogram

at a desired level: each connected connected component forms a cluster.

Material Selected from Introduction to Information Retrieval, Christopher D. Manning, Prabhakar Raghavan and Hinrich Schütze, 2008.

38

Page 39: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Hierarchical Complete-Linkage Clustering

• Similarity of the “furthest” points.• Makes “tighter,” spherical clusters that are

typically preferable.

Material Selected from Introduction to Information Retrieval, Christopher D. Manning, Prabhakar Raghavan and Hinrich Schütze, 2008.

39

Page 40: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Performance Optimization• 4 threads in Quad-Core system

Pre-Processing

Build Descriptor

Images Descriptors

Complete-Link Clustering

Calculate LBP

Build Descriptor

Compute all the similarities Clustering

40

Page 41: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Performance Optimization• 4 threads in Quad-Core system

– 3 times faster than single thread.• 73 minutes to 24 minute for 309 images.

Pre-Processing

Build Descriptor

Images Descriptors

Complete-Link Clustering

Calculate LBP

Build Descriptor

Compute all the similarities Clustering

41

Page 42: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

RESULT

42

Page 43: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

FERET Result• fa: gallery, 994 images• fb: alternative facial expression, 992 images• dup1: the photos taken after later, 736 images• dup2: the photos taken at least one year after

the gallery, 228 images

43

Page 44: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

FERET Result

44

Accuracy fb dup1 dup2LBP [PAMI2006] 95.67% 59.92% 45.61%

Our Result 98.89% 71.33% 68.42%

Time※ Register fb dup1 dup2LBP [PAMI2006] 46 seconds 102 seconds 67 seconds 21 seconds

Our Result 4 minutes 224 minutes 168 minutes 5 minutes

※ The time results are computed in multithreads version.

Page 45: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Experiments• Home Photo Dataset I

– 309 images, 5 subjects

• Home Photo Dataset II– 838 images, 8 subjects

45

Page 46: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Evaluation• Cluster Number• Unknown Number• Pair-wise Precision• Pair-wise Rand Index• Executing time

46

Page 47: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Unknown Number• The number of clusters which contain only one

component.

47

Page 48: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Precision/Rand Index

fptptpprecision

fntptprecall

tnfpfntptntpindexrand

48

Assigned

Same Cluster

Different Clusters

Ground Truth

Same Cluster tp fn

Different Clusters fp tn

Page 49: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Dataset-I Result#Clusters #Unknown Precision Rand

IndexTime ※

Picasa Web 94 73 99.92% 0.829816 10 seconds

Picasa PC 99 75 100% 0.865002 3 minutes

LBP [PAMI2006] 100 31 90.39% 0.811856 4 seconds

LID_PM [ICCV2009]

99 37 98.78% - 11 minutes

Our Result 100 39 99.46% 0.816290 24 minutes

LID+PM [Chang2010]

100 43 99.24% - 25 minutes

LBP: Local Binary Pattern, PM: Partial Matching, SB: Spatial Block

※ The time results are computed in multithreads version.49

Page 51: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Dataset-I Result• Wrong clustering result in LBP

• The clusters in our result

51

Page 52: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Dataset-II Result#Clusters #Unknown Precision Rand

IndexTime ※

Picasa Web 195 47 99.49% 0.876229 10 seconds

Picasa PC 253 150 100% 0.888492 10 minutes

LBP [PAMI2006] 253 45 91.70% 0.70261 17 seconds

LID_PM [ICCV2009] 253 79 97.88% - 51 minutes

Our Result 253 64 99.59% 0.871187 163 minutes

LID+PM [Chang2010]

253 89 99.82% - 136 minutes

LBP: Local Binary Pattern, PM: Partial Matching, SB: Spatial Block

※ The time results are computed in multithreads version.52

Page 54: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Dataset-II Result• Wrong clustering result in LBP

• Clusters in our result

54

Page 55: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Demo• Face Recognition with web camera

55

Page 56: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

CONCLUSION

56

Page 57: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Conclusion• LBP is an efficient algorithm for face

recognition.

• Partial Matching is good for the different facial expression or different illumination in facial images.

• Our system has better performance than LBP.

57

Page 58: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Future Works• Improve performance by GPU.

• Use other extension of LBP combined with Partial Matching.

58

Page 59: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Thank You for Your Attention!

59

Page 60: 以區域二元圖樣與部分比對為基礎之人臉辨識 Face Recognition with Local Binary Patterns  and Partial Matching

Result

LBP: Local Binary Pattern, PM: Partial Matching, SB: Spatial Block

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20.700000000000001

0.750000000000001

0.800000000000001

0.850000000000001

0.900000000000001

0.950000000000001

1

Home Photo Dataset I

LBP LBP+SB LBP+PM LBP+LID+SB

Recall

Prec

ision

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.20.700000000000001

0.750000000000001

0.800000000000001

0.850000000000001

0.900000000000001

0.950000000000001

1

Home Photo Dataset II

LBP LBP+SB LBP+PM LBP+SB+PM

Recall

Prec

ision

60