Search papers, labs, and topics across Lattice.
This paper introduces Human4K, a comprehensive 4K multi-view dataset designed to enhance whole-body 3D human reconstruction by addressing limitations in existing datasets. By integrating over six million high-resolution images with precise SMPL-X annotations from a synchronized motion capture system, the dataset captures complex and self-occluded body motions across diverse subjects. Experimental results demonstrate that models trained on Human4K achieve significant improvements in reconstruction accuracy, particularly for challenging areas like hands and feet under depth ambiguity.
Training on the Human4K dataset yields substantial gains in 3D human reconstruction accuracy, especially in depth-ambiguous scenarios.
Recent advances in 3D human reconstruction have improved overall performance, yet current models still fail in the most challenging real-world scenarios. They often produce unstable geometry, inaccurate limb articulation and unreliable predictions under depth ambiguity or self-occlusion. A key reason is that existing datasets still lack the combination of high-resolution images, high-precision annotations and diverse whole-body motions required to support robust reconstruction. To address this gap, we present Human4K, a large-scale 4K multi-view whole-body human reconstruction dataset with mocap-accurate SMPL-X annotations. Human4K contains over six million 4K images captured by an eight-view high-resolution camera system synchronized with a professional Vicon motion capture setup, covering 11 subjects performing complex, highly articulated and strongly self-occluded full-body motions. All sequences are processed by a Motion-Retargeting and Refinement Module (MRRM) to ensure precise alignment for the full body and extremities. Experimental results show that training with Human4K consistently improves whole-body reconstruction on standard benchmarks, with particularly large gains for hands, feet and depth-ambiguous limb configurations.
