Search papers, labs, and topics across Lattice.
This paper introduces Humanoid-OmniOcc, a comprehensive stereo-based occupancy dataset specifically designed for humanoid robots, addressing the limitations of existing datasets that are biased towards vehicle-centric perspectives. The dataset includes over 155K samples from 15 simulated indoor scenes and 5 real-world environments, facilitating the development of robust occupancy prediction models. Key experiments demonstrate that the proposed Humanoid-Surround-Stereo-guided Occupancy model significantly outperforms monocular baselines and effectively generalizes to both unseen simulated and real-world environments, showcasing the advantages of the Real2Sim2Real approach.
Humanoid-OmniOcc reveals that a stereo-based dataset can dramatically enhance occupancy prediction accuracy for humanoid robots, outperforming traditional monocular methods.
Occupancy prediction at voxel-level granularity is essential for safe robotic navigation and interaction in complex environments. Existing occupancy datasets, however, are predominantly designed for autonomous driving with vehicle-centric biases -- forward-facing cameras, far-field geometry, and static road priors -- limiting their applicability to embodied humanoid perception. We present Humanoid-OmniOcc, a large-scale panoramic stereo-based occupancy dataset tailored for humanoid robots. The dataset encompasses 15 diverse simulated indoor scenes and 5 real-world environments, yielding over 155K samples with broad scene and style diversity. Importantly, the dataset is designed around a Real2Sim2Real closed-loop paradigm: real sensor specifications drive physically accurate simulation, simulation produces large-scale annotated training data, and models trained in simulation are directly evaluated on real-world captures -- enabling iterative refinement of the sim-to-real pipeline. We further propose \textbf{H}umanoid \textbf{S}urround \textbf{S}tereo-guided \textbf{Occ}upancy model (Humanoid-OmniOcc) that exploits robust depth priors for accurate 2D-to-3D lifting. Extensive experiments show that Humanoid-OmniOcc consistently outperforms monocular baselines and generalizes well to both unseen simulated test scenes and real-world environments, validating the effectiveness of the Real2Sim2Real design. Code and data will be available upon acceptance at https://d-robotics-ai-lab.github.io/humanoid-omniocc.
