Multi-Expression Sequence Fusion Recognition Based on Probabilistic Cooperative Representation

Method	Test 1	Test 2	Test 3	Test 4	Test 5	Test 6	Test 7	Test 8	Test 9	Test 10	Average
Ours	91.1	88.4	86.7	87.7	88.0	87.1	91.2	86.5	88.1	90.3	88.5
Ori-AAM	67.7	71.2	73.9	71.5	71.9	70.1	68.1	73.3	72.9	72.3	71.3
Ori-pic	59.4	66.7	61.5	60.2	63.9	64.9	61.1	65.8	61.8	62.9	62.8

Table 2. Recognition rate in each state%
View table
Table 2. Recognition rate in each state%
Method 6-class 7-class
Related to people Not related to people Related to people Not related to people
Ours 99.4 88.5 99.0 83.2
Ori-AAM 92.9 71.3 89.5 68.5
Ori-pic 90.2 62.8 88.0 54.3

Table 3. Recognition rates of different methods related to person%

View table

Table 3. Recognition rates of different methods related to person%

Document	Method	6-class	7-class
Ref. [1]	STWLD+BHOF+SVM	91.6	—
Ref. [3]	DDL+CRC	99.5	98.9
Ref. [13]	5-frame CNN	92.9	—
Ref. [17]	STTM+SVM	97.7	—
Ref. [24]	ETEL_J	—	99.0
Ref. [25]	GLDPE	97.1	—
Ours	3-frame AAM+ProCRC	99.4	99.0

Table 4. Recognition rates of different documents unrelated to person
View table
Table 4. Recognition rates of different documents unrelated to person
Document Method 6-class 7-class
Ref. [23] CS-LSBP+HOAG+CCA+SVM — 81.4
Ref. [26] SLFDA 78.7 —
Ref. [27] AdaBoost+DBN 87.4 —
Ref. [28] WPHOG_TOP+SVM — 84.9
Ours 3-frame AAM+ProCRC 88.5 83.2

Tools

Get Citation

Copy Citation Text

Xiuyou Wang, Jianzhong Fan, Huaming Liu, Dongqing Xu, Zhengyan Liu. Multi-Expression Sequence Fusion Recognition Based on Probabilistic Cooperative Representation[J]. Laser & Optoelectronics Progress, 2019, 56(13): 131004

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites