Crowdsourcing facial expressions using popular gameplay

Crowdsourcing facial expressions using popular gameplay
复制标题

使用流行的游戏玩法众包面部表情

DOI:
10.1145/2542355.2542388
复制
发表时间:
2013
期刊:
SIGGRAPH Asia 2013 Technical Briefs
影响因子:
--
通讯作者:
Natalie Harrold
Natalie Harrold
中科院分区:
--
文献类型:
--
作者:
Chek Tien Tan;Daniel Rosser;Natalie Harrold

文献摘要

参考文献

被引文献

相似文献

面部表情分析系统通常采用机器学习算法,这些算法在很大程度上取决于训练它们的面部数据库的质量。不幸的是,生成高质量的人脸数据库是一项重大挑战,而且相当耗时。我们开发了 BeFaced,这是一款休闲平板电脑游戏,可实现大规模面部表情众包,以实现此类机器学习算法。基于流行的瓷砖匹配游戏机制,玩家需要在匹配的瓷砖上做出面部表情,以便清除它们并在游戏中前进。游戏中采用了识别精度的动态难度调整,以增加参与度,从而增加获得的各种面部表情的数量。每个面部表情都会被自动捕获、标记并发送到我们的在线面部数据库。在更抽象的层面上,BeFaced 研究了一种使用流行游戏机制来帮助计算机视觉算法进步的新颖方法。
Facial expression analysis systems often employ machine learning algorithms that depend a lot on the quality of the face database they are trained on. Unfortunately, generating high quality face databases is a major challenge that is rather time consuming. We have developed BeFaced, a tile-matching casual tablet game to enable massive crowdsourcing of facial expressions for the purpose of such machine learning algorithms. Based on the popular tile-matching gameplay mechanic, players are required to make facial expressions shown on matched tiles in order to clear them and advance in the game. Dynamic difficulty adjustment of the recognition accuracy is employed in the game in order to increase engagement and hence increase the quantity of varied facial expressions obtained. Each facial expression is automatically captured, labelled and sent to our online face database. At a more abstract level, BeFaced investigates a novel method of using popular game mechanics to aid the advancement of computer vision algorithms.
DOI: 10.4000/books.cedej.713
发表时间: 2021-09
期刊: Cross-Cultural Pragmatics
影响因子: --
作者:
Je suis d’ailleurs;Je suis d’ailleurs
通讯作者: Je suis d’ailleurs;Je suis d’ailleurs