Perceived emotions and AU combinations in ambiguous facial expressions

面部表情 心理学 计算机科学 人工智能 认知心理学 自然语言处理 模式识别(心理学) 语音识别 沟通 数学
作者
Wen-Jing Yan,Qian-Nan Ruan,Xiaolan Fu,Yuqi Sun
出处
期刊:Pattern Recognition Letters [Elsevier BV]
卷期号:164: 74-80 被引量:1
标识
DOI:10.1016/j.patrec.2022.10.018
摘要

• Network analysis depicts the emotion-network and the surprise is the most repulsive to other emotions. • AU25 and related mouth-open actions are the center of the AU-network. • Matrix calculation found that AU25 weights first in contributing to perceived emotion. • AU4 and AU25 are most frequently presented in the common ambiguous facial expressions Ambiguous facial expressions are common and cannot be classified as specifically prototypical or compound in nature. We have very little understanding of the relationships connecting such expressions to perceived human emotions, such as in terms of compatibility or repulsion. This ignorance also exists with regards to their relationship to action units (AUs). This research employed network analysis to depict the network of perceived emotions and AUs in nearly 5,000 facial expressions obtained from the RAF-AU database, and calculated the centrality indices. We then used a matrix calculation to analyze the relationships between AU combinations and perceived emotions to better understand how people interpret the actions appearing on faces. The results showed that: (1) surprise was the most repulsive to other emotions in the emotion network, (2) AU25 and related open-mouth actions comprised the center of the AU network, (3) AU25 was weighted first in terms of contributing to perceived emotion, and (4) AU4 and AU25 were the most frequently presented in common ambiguous facial expressions. The results were not consistent with those of previous research, mainly due to differences in research methods and materials. The results imply that emotions perceived from ambiguous facial expressions cannot be predicted by core AUs of prototypical facial expressions. The implications and limitations of these conclusions are also discussed herein.

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
SevenKing完成签到,获得积分20
刚刚
刚刚
2秒前
英俊的铭应助www采纳,获得10
2秒前
xx完成签到,获得积分20
2秒前
2秒前
SevenKing发布了新的文献求助10
2秒前
2秒前
FashionBoy应助lhy123采纳,获得10
3秒前
zzzkk完成签到,获得积分10
3秒前
小狮子发布了新的文献求助10
6秒前
乐观的问枫完成签到 ,获得积分10
6秒前
6秒前
7秒前
小七完成签到,获得积分20
7秒前
Pessica发布了新的文献求助10
7秒前
9秒前
9秒前
研友_aLjNNL完成签到,获得积分10
9秒前
10秒前
10秒前
Sirius发布了新的文献求助10
11秒前
放逐心灵完成签到,获得积分20
11秒前
知识四面八方来完成签到 ,获得积分10
12秒前
12秒前
Khan发布了新的文献求助10
12秒前
伊绵好完成签到,获得积分10
12秒前
13秒前
碎片发布了新的文献求助10
13秒前
dddlll发布了新的文献求助10
14秒前
skyler发布了新的文献求助10
14秒前
小蘑菇应助suuu采纳,获得30
14秒前
zizizizi发布了新的文献求助10
15秒前
TEE发布了新的文献求助10
16秒前
无极微光应助如意的书白采纳,获得20
16秒前
活泼学生发布了新的文献求助10
16秒前
小碗发布了新的文献求助10
17秒前
Alvaro发布了新的文献求助10
17秒前
18秒前
19秒前
高分求助中
Principles of Economics, 11th Edition 10000
University Physics with Modern Physics, 16th edition 10000
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
Matrix Methods in Data Mining and Pattern Recognition 510
Reading and Understanding Health Research 500
Social Skills Improvement System-Rating Scales--Chinese Version 500
Dynamische Polarisation von H-1 und B-11 in (CH-3)-3NBH-3 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 工程类 有机化学 化学工程 生物化学 计算机科学 内科学 物理 复合材料 催化作用 细胞生物学 无机化学 光电子学 物理化学 电极 基因
热门帖子
关注 科研通微信公众号,转发送积分 7251549
求助须知:如何正确求助?哪些是违规求助? 8874035
关于积分的说明 18730628
捐赠科研通 6931418
什么是DOI,文献DOI怎么找? 3199473
关于科研通互助平台的介绍 2374329
邀请新用户注册赠送积分活动 2174053