Replica Dataset is a high-quality 3D dataset of realistic indoor environments designed to advance research in computer vision, robotics, and embodied AI. Developed by Facebook Research (now Meta AI), it features accurate geometric reconstructions, high-resolution and high dynamic range textures, and comprehensive semantic annotations. Each environment contains detailed models of real-world spaces, including rooms, furniture, glass, and mirror surfaces. The dataset also provides semantic and instance segmentations, planar decomposition, and navigation meshes, making it highly suitable for simulation, visual perception, and autonomous navigation tasks.
Replica integrates seamlessly with AI Habitat, Meta’s framework for embodied AI training, enabling large-scale agent simulation and photorealistic rendering for reinforcement learning and robotics. Researchers can use Replica’s ReplicaViewer to interactively explore the 3D scenes.
Features
- Contains 18 high-quality indoor 3D scenes with dense geometry and HDR textures
- Includes semantic and instance segmentation, planar decomposition, and navigation meshes
- Offers ReplicaViewer for interactive exploration and ReplicaRenderer for automated rendering
- Fully compatible with AI Habitat for embodied AI research and agent training
- Provides glass and mirror surface parameterization for advanced optical realism
- Designed for use in visual perception, robotics, AR/VR, and simulation research
Project Samples
