ERIT
收藏资源简介:
ERIT数据集由香港科技大学电子与计算机工程系创建,旨在促进轻量级多模态融合研究。该数据集包含来自老年人视频的文本和图像数据,每个样本带有七种情绪标签。数据集内容丰富,包括多种情绪表达和个体差异,适用于训练和评估老年人情绪识别模型。数据集的创建过程包括从视频中提取帧、使用自动语音识别进行文本转录以及通过DeepFace框架进行情绪标签验证。ERIT数据集主要应用于老年人情绪识别、医疗保健和辅助技术等领域,旨在解决老年人情绪识别的准确性和鲁棒性问题。
The ERIT dataset was created by the Department of Electronic and Computer Engineering at The Hong Kong University of Science and Technology, aiming to promote research on lightweight multimodal fusion. This dataset includes text and image data sourced from videos of elderly individuals, with each sample annotated with seven emotion labels. Featuring rich content covering diverse emotional expressions and individual differences, the dataset is applicable for training and evaluating emotion recognition models for elderly populations. The dataset construction process involves three main procedures: extracting frames from the original videos, conducting text transcription via automatic speech recognition (ASR), and verifying emotion labels using the DeepFace framework. The ERIT dataset is mainly applied in fields including elderly emotion recognition, healthcare, and assistive technologies, with the goal of resolving the issues of accuracy and robustness in elderly emotion recognition.




