MuSe-Sent: Multimodal Sentiment Classification in-the-Wild (MuSe2021)
收藏资源简介：
MuSe-Sent of the 2nd Multimodal Sentiment in-the-Wild Challenge! Predicting five advanced intensity classes for each of the emotional dimensions (valence, arousal) for segments of audio-video-text data. This package includes only MuSe-Sent features (all partitions) and labels of the training and development set (test scoring via the MuSe website). More: https://www.muse-challenge.org/muse2021 General: The purpose of the Multimodal Sentiment Analysis in Real-life media Challenge and Workshop (MuSe) is to bring together communities from different disciplines. We introduce the novel dataset MuSe-CAR that covers the range of aforementioned desiderata. MuSe-CAR is a large (>36h), multimodal dataset which has been gathered in-the-wild with the intention of further understanding Multimodal Sentiment Analysis in-the-wild, e.g., the emotional engagement that takes place during product reviews (i.e., automobile reviews) where a sentiment is linked to a topic or entity. We have designed MuSe-CAR to be of high voice and video quality, as informative video social media content, as well as everyday recording devices have improved in recent years. This enables robust learning, even with a high degree of novel, in-the-wild characteristics, for example as related to: i) Video: Shot size (a mix of close-up, medium, and long shots), face-angle (side, eye, low, high), camera motion (free, free but stable, and free but unstable, switch, e.g., zoom, fixed), reviewer visibility (full body, half-body, face only, and hands only), highly varying backgrounds, and people interacting with objects (car parts). ii) Audio: Ambient noises (car noises, music), narrator and host diarisation, diverse microphone types, and speaker locations. iii) Text: Colloquialisms, and domain-specific terms.
第二届野外多模态情感分析挑战赛（2nd Multimodal Sentiment in-the-Wild Challenge）的MuSe-Sent赛道任务：针对音视频文本（audio-video-text）数据片段，为情感维度（效价valence、唤醒度arousal）分别预测五个进阶强度等级。本数据包仅包含MuSe-Sent特征（全分区）以及训练集与开发集的标签，测试集评分需通过MuSe官方网站提交获取。更多详情：https://www.muse-challenge.org/muse2021 通用说明：现实媒体多模态情感分析挑战赛与研讨会（Multimodal Sentiment Analysis in Real-life media Challenge and Workshop，简称MuSe）的宗旨是汇聚不同学科领域的研究者群体。本次挑战赛推出了全新数据集MuSe-CAR，其覆盖了前述各项研究目标。 MuSe-CAR是一个时长超过36小时的大型多模态数据集，采集自真实野外场景，旨在深化对野外多模态情感分析的研究——例如汽车评测类产品评论中的情感参与度分析，这类场景中情感表达与特定主题或实体绑定。得益于近年来社交媒体高信息量视频内容与日常录制设备的品质升级，我们将MuSe-CAR设计为具备高音质与视频画质的数据集，这使得模型即便面对大量新颖的野外场景特征，仍可实现稳健的模型训练。相关特征示例包括： i) 视频维度：镜头景别（特写、中景、远景混合）、面部角度（侧面、眼部特写、低角度、高角度）、相机运动（自由移动、稳定自由移动、不稳定自由移动、变焦、固定）、拍摄者出镜状态（全身、半身、仅面部、仅手部）、高度多变的背景，以及人与汽车零部件等物体的交互行为。 ii) 音频维度：环境噪声（汽车噪音、背景音乐）、旁白与主持者的说话人分段标注、多样的麦克风类型，以及不同的说话人位置。 iii) 文本维度：口语化表达与领域专属术语。
