SPOT: SE(3) Pose Trajectory Diffusion for Object-Centric Manipulation
On the map: Offline RL and imitation learning from data → Robot manipulation imitation learning
Abstract (arXiv)
We introduce SPOT, an object-centric imitation learning framework. The key idea is to capture each task by an object-centric representation, specifically the SE(3) object pose trajectory relative to the target. This approach decouples embodiment actions from sensory inputs, facilitating learning from various demonstration types, including both action-based and action-less human hand demonstrations, as well as cross-embodiment generalization. Additionally, object pose trajectories inherently capture planning constraints from demonstrations without the need for manually-crafted rules. To guide the robot in executing the task, the object trajectory is used to condition a diffusion policy. We systematically evaluate our method on simulation and real-world tasks. In real-world evaluation, using only eight demonstrations shot on an iPhone, our approach completed all tasks while fully complying with task constraints. Project page: https://nvlabs.github.io/object_centric_diffusion
Key facts. Cited by 36 papers in the map (data through 7 Oct 2026). Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data. First posted on arXiv 1 Nov 2024.
Builds on
No reference list yet.
Built on by
Citing papers per year: 2025: 19; 2026: 17.
36 papers.
Sort:
By area:
- 2025UniVLA: Learning to Act Anywhere with Task-centric Latent Actionscited by 386
- 2025Humanoid Policy ~ Human Policycited by 65
- 2025ControlVLA: Few-shot Object-centric Adaptation for Pre-trained Vision-Language-Action Modelscited by 43
- 2025Robotic Manipulation by Imitating Generated Videos Without Physical Demonstrationscited by 33
- 2026CLAP: Contrastive Latent Action Pretraining for Learning Vision-Language-Action Models from Human Videoscited by 27
- 2025Robot Learning from a Physical World Modelcited by 18
- 2026HumanEgo: Zero-Shot Robot Learning from Minutes of Human Egocentric Videoscited by 14
- 2025Object-centric 3D Motion Field for Robot Learning from Human Videoscited by 13
- 2026SimToolReal: An Object-Centric Policy for Zero-Shot Dexterous Tool Manipulationcited by 10
- 2025GenFlowRL: Shaping Rewards with Generative Object-Centric Flow in Visual Reinforcement Learningcited by 8
- 2026GuidedVLA: Specifying Task-Relevant Factors via Plug-and-Play Action Attention Specializationcited by 7
- 2026FlowHOI: Flow-based Semantics-Grounded Generation of Hand-Object Interactions for Dexterous Robot Manipulationcited by 6
24 more
- 2026Learning Dexterous Manipulation Using Contact Wrench Guidance From Human Demonstrationcited by 6
- 2025PRISM-DP: Spatial Pose-based Observations for Diffusion-Policies via Segmentation, Mesh Generation, and Pose Trackingcited by 5
- 2026OCRA: Object-Centric Learning with 3D and Tactile Priors for Human-to-Robot Action Transfercited by 5
- 2025V-HOP: Visuo-Haptic 6D Object Pose Trackingcited by 4
- 2025BoxDreamer: Dreaming Box Corners for Generalizable Object Pose Estimationcited by 3
- 2025FUNCanon: Learning Pose-Aware Action Primitives via Functional Object Canonicalization for Generalizable Robotic Manipulationcited by 3
- 2025AFFORD2ACT: Affordance-Guided Automatic Keypoint Selection for Generalizable and Lightweight Robotic Manipulationcited by 3
- 2025Multimodal Diffusion Forcing for Forceful Manipulationcited by 3
- 2025Knowledge-Driven Imitation Learning: Enabling Generalization Across Diverse Conditionscited by 2
- 2025DeLTa: Demonstration and Language-Guided Novel Transparent Object Manipulationcited by 2
- 2025Human Centric General Physical Intelligence for Agile Manufacturing Automationcited by 1
- 2026Imitating What Works: Simulation-Filtered Modular Policy Learning from Human Videoscited by 1
- 2025Imitation Learning with Precisely Labeled Human Demonstrationscited by 0
- 2025ReSeFlow: Rectifying SE(3)-Equivariant Policy Learning Flowscited by 0
- 2025GLUE: Global-Local Unified Encoding for Imitation Learning via Key-Patch Trackingcited by 0
- 2026EgoFlow: Gradient-Guided Flow Matching for Egocentric 6DoF Object Motion Generationcited by 0
- 2026Spatially Conditioned Diffusion Policy: Learning Precise and Robust Manipulation with a Single RGB Cameracited by 0
- 2026Human2Any: Human-to-Robot Transfer via Constraint-Aware Compositional Planningcited by 0
- 2026EmbodimentSemantic: A Spatial Scene-Graph Dataset and Benchmark for Vision-Language Models on Embodied Manipulation Trajectoriescited by 0
- 2026The Embodiment Gap in Robot Foundation Modelscited by 0
- 2026RoboTok: A Scalable Data Engine for Internet Demonstration Video Retrieval and Dexterous Manipulation Learningcited by 0
- 2026GeomVLA: Unifying Scene, Motion, and Action in 3Dcited by 0
- 2026From Pixel to Poses: Object-centric Tool Manipulation Learning from Human Demonstrationscited by 0
- 2026World Motion Models: Flexible Sequence Modeling of SE(3) Trajectoriescited by 0
- 2026World Motion Models: Flexible Sequence Modeling of SE(3) Trajectories1 Oct 2026
- 2026From Pixel to Poses: Object-centric Tool Manipulation Learning from Human Demonstrations28 Sep 2026
- 2026GeomVLA: Unifying Scene, Motion, and Action in 3D12 Sep 2026
- 2026RoboTok: A Scalable Data Engine for Internet Demonstration Video Retrieval and Dexterous Manipulation Learning2 Sep 2026
- 2026The Embodiment Gap in Robot Foundation Models19 Aug 2026
- 2026Human2Any: Human-to-Robot Transfer via Constraint-Aware Compositional Planning27 Jun 2026
- 2026Learning Dexterous Manipulation Using Contact Wrench Guidance From Human Demonstration22 Jun 2026
- 2026Spatially Conditioned Diffusion Policy: Learning Precise and Robust Manipulation with a Single RGB Camera12 Jun 2026
- 2026EmbodimentSemantic: A Spatial Scene-Graph Dataset and Benchmark for Vision-Language Models on Embodied Manipulation Trajectories6 Jun 2026
- 2026HumanEgo: Zero-Shot Robot Learning from Minutes of Human Egocentric Videos24 May 2026
- 2026GuidedVLA: Specifying Task-Relevant Factors via Plug-and-Play Action Attention Specialization12 May 2026
- 2026EgoFlow: Gradient-Guided Flow Matching for Egocentric 6DoF Object Motion Generation1 Apr 2026
24 more
- 2026OCRA: Object-Centric Learning with 3D and Tactile Priors for Human-to-Robot Action Transfer15 Mar 2026
- 2026SimToolReal: An Object-Centric Policy for Zero-Shot Dexterous Tool Manipulation18 Feb 2026
- 2026FlowHOI: Flow-based Semantics-Grounded Generation of Hand-Object Interactions for Dexterous Robot Manipulation13 Feb 2026
- 2026Imitating What Works: Simulation-Filtered Modular Policy Learning from Human Videos13 Feb 2026
- 2026CLAP: Contrastive Latent Action Pretraining for Learning Vision-Language-Action Models from Human Videos7 Jan 2026
- 2025Robot Learning from a Physical World Model10 Nov 2025
- 2025Multimodal Diffusion Forcing for Forceful Manipulation6 Nov 2025
- 2025DeLTa: Demonstration and Language-Guided Novel Transparent Object Manipulation7 Oct 2025
- 2025AFFORD2ACT: Affordance-Guided Automatic Keypoint Selection for Generalizable and Lightweight Robotic Manipulation1 Oct 2025
- 2025GLUE: Global-Local Unified Encoding for Imitation Learning via Key-Patch Tracking27 Sep 2025
- 2025FUNCanon: Learning Pose-Aware Action Primitives via Functional Object Canonicalization for Generalizable Robotic Manipulation23 Sep 2025
- 2025ReSeFlow: Rectifying SE(3)-Equivariant Policy Learning Flows20 Sep 2025
- 2025Human Centric General Physical Intelligence for Agile Manufacturing Automation16 Aug 2025
- 2025GenFlowRL: Shaping Rewards with Generative Object-Centric Flow in Visual Reinforcement Learning14 Aug 2025
- 2025Robotic Manipulation by Imitating Generated Videos Without Physical Demonstrations1 Jul 2025
- 2025Knowledge-Driven Imitation Learning: Enabling Generalization Across Diverse Conditions26 Jun 2025
- 2025ControlVLA: Few-shot Object-centric Adaptation for Pre-trained Vision-Language-Action Models19 Jun 2025
- 2025Object-centric 3D Motion Field for Robot Learning from Human Videos4 Jun 2025
- 2025UniVLA: Learning to Act Anywhere with Task-centric Latent Actions9 May 2025
- 2025PRISM-DP: Spatial Pose-based Observations for Diffusion-Policies via Segmentation, Mesh Generation, and Pose Tracking29 Apr 2025
- 2025Imitation Learning with Precisely Labeled Human Demonstrations18 Apr 2025
- 2025BoxDreamer: Dreaming Box Corners for Generalizable Object Pose Estimation10 Apr 2025
- 2025Humanoid Policy ~ Human Policy17 Mar 2025
- 2025V-HOP: Visuo-Haptic 6D Object Pose Tracking24 Feb 2025
- 2025UniVLA: Learning to Act Anywhere with Task-centric Latent Actionscited by 386
- 2025ControlVLA: Few-shot Object-centric Adaptation for Pre-trained Vision-Language-Action Modelscited by 43
- 2026CLAP: Contrastive Latent Action Pretraining for Learning Vision-Language-Action Models from Human Videoscited by 27
- 2025Human Centric General Physical Intelligence for Agile Manufacturing Automationcited by 1
- 2026The Embodiment Gap in Robot Foundation Modelscited by 0
- 2026GeomVLA: Unifying Scene, Motion, and Action in 3Dcited by 0
- 2026GeomVLA: Unifying Scene, Motion, and Action in 3D12 Sep 2026
- 2026The Embodiment Gap in Robot Foundation Models19 Aug 2026
- 2026CLAP: Contrastive Latent Action Pretraining for Learning Vision-Language-Action Models from Human Videos7 Jan 2026
- 2025Human Centric General Physical Intelligence for Agile Manufacturing Automation16 Aug 2025
- 2025ControlVLA: Few-shot Object-centric Adaptation for Pre-trained Vision-Language-Action Models19 Jun 2025
- 2025UniVLA: Learning to Act Anywhere with Task-centric Latent Actions9 May 2025
- 2025Multimodal Diffusion Forcing for Forceful Manipulationcited by 3
- 2025Knowledge-Driven Imitation Learning: Enabling Generalization Across Diverse Conditionscited by 2
- 2025Imitation Learning with Precisely Labeled Human Demonstrationscited by 0
- 2025GLUE: Global-Local Unified Encoding for Imitation Learning via Key-Patch Trackingcited by 0
- 2026From Pixel to Poses: Object-centric Tool Manipulation Learning from Human Demonstrationscited by 0
- 2026From Pixel to Poses: Object-centric Tool Manipulation Learning from Human Demonstrations28 Sep 2026
- 2025Multimodal Diffusion Forcing for Forceful Manipulation6 Nov 2025
- 2025GLUE: Global-Local Unified Encoding for Imitation Learning via Key-Patch Tracking27 Sep 2025
- 2025Knowledge-Driven Imitation Learning: Enabling Generalization Across Diverse Conditions26 Jun 2025
- 2025Imitation Learning with Precisely Labeled Human Demonstrations18 Apr 2025
- 2025PRISM-DP: Spatial Pose-based Observations for Diffusion-Policies via Segmentation, Mesh Generation, and Pose Trackingcited by 5
- 2025ReSeFlow: Rectifying SE(3)-Equivariant Policy Learning Flowscited by 0
- 2026EgoFlow: Gradient-Guided Flow Matching for Egocentric 6DoF Object Motion Generationcited by 0
- 2026Spatially Conditioned Diffusion Policy: Learning Precise and Robust Manipulation with a Single RGB Cameracited by 0
- 2026Spatially Conditioned Diffusion Policy: Learning Precise and Robust Manipulation with a Single RGB Camera12 Jun 2026
- 2026EgoFlow: Gradient-Guided Flow Matching for Egocentric 6DoF Object Motion Generation1 Apr 2026
- 2025ReSeFlow: Rectifying SE(3)-Equivariant Policy Learning Flows20 Sep 2025
- 2025PRISM-DP: Spatial Pose-based Observations for Diffusion-Policies via Segmentation, Mesh Generation, and Pose Tracking29 Apr 2025
- 2025Robot Learning from a Physical World Modelcited by 18
- 2025Robot Learning from a Physical World Model10 Nov 2025
- 2025Humanoid Policy ~ Human Policycited by 65
- 2025Robotic Manipulation by Imitating Generated Videos Without Physical Demonstrationscited by 33
- 2026HumanEgo: Zero-Shot Robot Learning from Minutes of Human Egocentric Videoscited by 14
- 2025Object-centric 3D Motion Field for Robot Learning from Human Videoscited by 13
- 2026SimToolReal: An Object-Centric Policy for Zero-Shot Dexterous Tool Manipulationcited by 10
- 2025GenFlowRL: Shaping Rewards with Generative Object-Centric Flow in Visual Reinforcement Learningcited by 8
- 2026FlowHOI: Flow-based Semantics-Grounded Generation of Hand-Object Interactions for Dexterous Robot Manipulationcited by 6
- 2026Learning Dexterous Manipulation Using Contact Wrench Guidance From Human Demonstrationcited by 6
- 2026OCRA: Object-Centric Learning with 3D and Tactile Priors for Human-to-Robot Action Transfercited by 5
- 2025V-HOP: Visuo-Haptic 6D Object Pose Trackingcited by 4
- 2025BoxDreamer: Dreaming Box Corners for Generalizable Object Pose Estimationcited by 3
- 2025FUNCanon: Learning Pose-Aware Action Primitives via Functional Object Canonicalization for Generalizable Robotic Manipulationcited by 3
12 most cited of 18 All 18 (JSON)
- 2026World Motion Models: Flexible Sequence Modeling of SE(3) Trajectories1 Oct 2026
- 2026RoboTok: A Scalable Data Engine for Internet Demonstration Video Retrieval and Dexterous Manipulation Learning2 Sep 2026
- 2026Human2Any: Human-to-Robot Transfer via Constraint-Aware Compositional Planning27 Jun 2026
- 2026Learning Dexterous Manipulation Using Contact Wrench Guidance From Human Demonstration22 Jun 2026
- 2026HumanEgo: Zero-Shot Robot Learning from Minutes of Human Egocentric Videos24 May 2026
- 2026OCRA: Object-Centric Learning with 3D and Tactile Priors for Human-to-Robot Action Transfer15 Mar 2026
- 2026SimToolReal: An Object-Centric Policy for Zero-Shot Dexterous Tool Manipulation18 Feb 2026
- 2026FlowHOI: Flow-based Semantics-Grounded Generation of Hand-Object Interactions for Dexterous Robot Manipulation13 Feb 2026
- 2026Imitating What Works: Simulation-Filtered Modular Policy Learning from Human Videos13 Feb 2026
- 2025DeLTa: Demonstration and Language-Guided Novel Transparent Object Manipulation7 Oct 2025
- 2025AFFORD2ACT: Affordance-Guided Automatic Keypoint Selection for Generalizable and Lightweight Robotic Manipulation1 Oct 2025
- 2025FUNCanon: Learning Pose-Aware Action Primitives via Functional Object Canonicalization for Generalizable Robotic Manipulation23 Sep 2025
12 newest of 18 All 18 (JSON)
Similar, not cited
Close in wording, no citation either way
- 2024Motion Before Action: Diffusing Object Motion as Manipulation Condition
- 2026Hand-centric Human-to-Robot Trajectory Transfer from Video Demonstrations via Object Category-agnostic Temporal Localization
- 2024Language-Guided Object-Centric Diffusion Policy for Generalizable and Collision-Aware Robotic Manipulation
- 2026RE4: Transformation-aware Imitation of Object Interactions Using Manipulation Modes
- 2025X-Diffusion: Training Diffusion Policies on Cross-Embodiment Human Demonstrations
- 2023One-Shot Imitation Learning: A Pose Estimation Perspective
- 2026More Structure, Not More Capacity: Object-Centric Representations for Visuomotor Imitation Learning
- 2025DemoDiffusion: One-Shot Human Imitation using pre-trained Diffusion Policy