HUMAN4D

HUMAN4D 是一个大型的多模式 4D 数据集，其中包含由基于标记的专业 MoCap、体积捕获和音频记录系统同时捕获的各种人类活动。通过捕捉 2 位女性和 2 美元的男性专业演员表演各种全身动作和表情，HUMAN4D 提供了在单人和多人日常、身体和社交活动（跳跃、跳舞等）中遇到的各种动作和姿势. )，以及多 RGBD (mRGBD)、体积和音频数据。描述来自：HUMAN4D：运动和沉浸式媒体的以人为中心的多模态数据集

HUMAN4D is a large-scale multimodal 4D dataset that encompasses a diverse array of human activities simultaneously captured via marker-based professional MoCap, volumetric capture, and audio recording systems. By having two female and two male professional actors perform a wide variety of full-body movements and expressions, HUMAN4D provides various motions and poses encountered in single- and multi-person daily, physical and social activities such as jumping, dancing, and other scenarios, alongside multi-RGBD (mRGBD), volumetric, and audio data. Adapted from: HUMAN4D: A Human-Centric Multimodal Dataset for Motions and Immersive Media

搜集汇总

背景与挑战

背景概述

HUMAN4D是一个大型多模式4D数据集，包含专业演员表演的各种全身动作和表情，涵盖单人和多人日常、身体和社交活动，并提供多RGBD、体积和音频数据。