CONVERSE

CONVERSE数据集是由斯旺西大学计算机科学与医学图像分析组创建的，专注于复杂对话交互的人体动作识别。该数据集包含8个对话场景，通过两个Kinect深度传感器收集，旨在提供一个包含多种基本动作、交互和运动的事件集合，以更真实地代表现实世界中的交互。数据集的应用领域主要集中在通过3D人体姿势特征进行对话交互分类的研究，旨在解决如何从复杂的对话交互中识别和理解人体动作的问题。

The CONVERSE dataset was created by the Computer Science and Medical Image Analysis Group of Swansea University, focusing on human action recognition in complex conversational interactions. This dataset includes 8 conversational scenarios, collected using two Kinect depth sensors. It aims to provide a collection of events containing various basic actions, interactions and movements, to more realistically represent real-world interactions. The main application fields of the dataset are research on conversational interaction classification via 3D human pose features, with the goal of solving the problem of how to recognize and understand human actions from complex conversational interactions.