ECEM

该数据集名为ECEM，它将视频片段与角色情绪的详细解释相结合，以便深入探讨多模态对话中情绪表达背后的因果因素。该数据集专注于非中性情绪，并采用了一套详细的标注方法以确保标注的准确性和一致性。它的目标是生成多种情绪的多个因果解释，旨在支持多模态对话情绪原因解释任务（Mcece）。This dataset, named ECEM, combines video clips with detailed explanations of characters' emotions to enable an in-depth exploration of the causal factors behind emotional expressions in multimodal conversations. It focuses on non-neutral emotions and employs a rigorous annotation protocol to guarantee the accuracy and consistency of its annotations. The goal of this dataset is to generate multiple causal explanations for a range of emotions, aiming to support the Multimodal Conversation Emotion Cause Explanation (Mcece) task.