The Indonesian Young-Adult Attachment (IYAA) Dataset: A Multimodal Behavioral Young-Adult Attachment Assessment

IYAA is an Indonesian language-based facial expression and speech for attachment style classification dataset. IYAA consists of Indonesian young-adult audio-video data from diverse age groups between 18-29 years old, with 20 male and 67 female subjects. The data was developed during the COVID-19 pandemic in an online setting by using the Zoom video conference application. Each subject participated in a projective-based attachment assessment which consists of two stages: exposure and response of 14 attachment-based stimuli. The data obtained for each subject includes 14 exposure videos, 14 response videos, and 14 speech response audios. However, two subjects' data are incomplete due to technical issues during data collection. The first subject loses an exposure video, and the second one loses data for a stimulus (an exposure video, a response video, and a speech response audio). So, in total, there are 3650 files, including 1216 exposure videos, 1217 response videos, and 1217 speech response audios. Each data has a varying duration; for exposure video, the minimum duration is 30 seconds, while for response video and speech response audio, the minimum duration is 40 seconds. The data is annotated into two kinds of annotation: attachment and emotion. First, each subject is annotated into one label among three classes of attachment style: secure, insecure anxious, insecure avoidance. Second, each stimulus's data (exposure videos, response videos, and speech response audios) are annotated into one or more labels among eight basic emotions: neutral, happy, sad, contempt, anger, disgust, surprise, and fear. Subject's responses on ECR-RS self-report questionnaire are also extracted to support holistic adult attachment assessment regarding non-verbal behaviors and subjective responses. The IYAA dataset supports multidisciplinary research, specifically on attachment, in Psychology (Behavioral, Social, and Developmental Psychology) and Computer Science (Social Signal Processing).IYAA是一款面向依恋风格（attachment style）分类任务、基于印尼语的面部表情与语音数据集。该数据集涵盖18至29岁不同年龄段的印尼青年成年人群体的音视频数据，共纳入20名男性与67名女性受试者。 数据采集工作于新冠疫情期间通过Zoom视频会议应用以线上形式开展。每名受试者参与了基于投射测验的依恋评估，该评估分为两个阶段：14项依恋相关刺激的呈现与回应环节。 每名受试者的采集数据包含14条呈现视频、14条回应视频与14条语音回应音频。不过受数据采集期间的技术故障影响，有两名受试者的数据存在缺失：第一名受试者丢失了1条呈现视频，第二名受试者丢失了某一刺激对应的全部数据（含1条呈现视频、1条回应视频与1条语音回应音频）。因此数据集总计包含3650个文件，其中包括1216条呈现视频、1217条回应视频与1217条语音回应音频。 各类数据的时长存在差异：呈现视频的最短时长为30秒，回应视频与语音回应音频的最短时长均为40秒。 该数据集包含两类标注体系：依恋风格标注与情绪标注。其一，每名受试者被标注为三类依恋风格中的一类：安全型、不安全焦虑型、不安全回避型。其二，每项刺激对应的数据（呈现视频、回应视频与语音回应音频）被标注为八种基本情绪中的一种或多种：中性、快乐、悲伤、轻蔑、愤怒、厌恶、惊讶与恐惧。 研究人员还提取了受试者填写的ECR-RS自我报告问卷（Experiences in Close Relationships-Revised Scale）结果，以辅助开展针对非言语行为与主观回应的全方位成人依恋评估。 IYAA数据集可支撑多学科领域的相关研究，尤其可服务于心理学（行为心理学、社会心理学与发展心理学）及计算机科学（社会信号处理（Social Signal Processing））领域中围绕依恋主题展开的研究。

搜集汇总

背景与挑战

背景概述

该数据集是一个用于依恋风格分类的印尼年轻成人多模态行为数据集，包含87名18-29岁参与者的音频-视频数据，总计3650个文件，数据收集于COVID-19疫情期间的在线环境。数据集标注了依恋风格（安全、不安全焦虑、不安全回避）和八种基本情绪，支持心理学和计算机科学领域的多学科研究，但访问受限，仅限学术用途。