PocketDP3: Efficient Pocket-Scale 3D Visuomotor Policy

• URL: http://arxiv.org/abs/2601.22018v2
• Date: Fri, 30 Jan 2026 13:52:41 GMT
• Title: PocketDP3: Efficient Pocket-Scale 3D Visuomotor Policy
• Abstract summary: 3D vision-based diffusion policies have shown strong capability in learning complex robotic manipulation skills.<n>A common architectural mismatch exists in these models: a tiny yet efficient point-cloud encoder is often paired with a massive decoder.<n>We propose PocketDP3, a pocket-scale 3D diffusion policy that replaces the heavy conditional U-Net decoder.
• Score: 2.661051697105248
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Recently, 3D vision-based diffusion policies have shown strong capability in learning complex robotic manipulation skills. However, a common architectural mismatch exists in these models: a tiny yet efficient point-cloud encoder is often paired with a massive decoder. Given a compact scene representation, we argue that this may lead to substantial parameter waste in the decoder. Motivated by this observation, we propose PocketDP3, a pocket-scale 3D diffusion policy that replaces the heavy conditional U-Net decoder used in prior methods with a lightweight Diffusion Mixer (DiM) built on MLP-Mixer blocks. This architecture enables efficient fusion across temporal and channel dimensions, significantly reducing model size. Notably, without any additional consistency distillation techniques, our method supports two-step inference without sacrificing performance, improving practicality for real-time deployment. Across three simulation benchmarks--RoboTwin2.0, Adroit, and MetaWorld--PocketDP3 achieves state-of-the-art performance with fewer than 1% of the parameters of prior methods, while also accelerating inference. Real-world experiments further demonstrate the practicality and transferability of our method in real-world settings. Code will be released.

Related papers

• Hydra-DP3: Frequency-Aware Right-Sizing of 3D Diffusion Policies for Visuomotor Control [2.661051697105248]
  Diffusion-based visuomotor policies perform well in robotic manipulation.<n>Current methods still inherit image-generation-style decoders and multi-step sampling.<n>We propose Hydra-DP3 (HDP3), a pocket-scale 3D diffusion policy with a lightweight Diffusion Mixer decoder.
  arXiv  Detail & Related papers  (2026-05-02T19:07:09Z)
• Efficient Encoder-Free Fourier-based 3D Large Multimodal Model [22.758597018527244]
  Large Multimodal Models (LMMs) that process 3D data typically rely on heavy, pre-trained visual encoders to extract geometric features.<n>We propose Fase3D, the first efficient encoder-free 3D scene LMM.
  arXiv  Detail & Related papers  (2026-02-26T16:16:02Z)
• Information Filtering via Variational Regularization for Robot Manipulation [2.564607372946005]
  We propose a lightweight module that imposes a timestep-conditioned Gaussian over backbone features and applies a KL-divergence regularizer.<n>Our approach improves the success rate by 6.1% on RoboTwin2.0 and by 4.1% on Adroit and MetaWorld, achieving new state-of-the-art results.
  arXiv  Detail & Related papers  (2026-01-29T16:17:42Z)
• TriCLIP-3D: A Unified Parameter-Efficient Framework for Tri-Modal 3D Visual Grounding based on CLIP [52.79100775328595]
  3D visual grounding allows an embodied agent to understand visual information in real-world 3D environments based on human instructions.<n>Existing 3D visual grounding methods rely on separate encoders for different modalities.<n>We propose a unified 2D pre-trained multi-modal network to process all three modalities.
  arXiv  Detail & Related papers  (2025-07-20T10:28:06Z)
• Rapid Salient Object Detection with Difference Convolutional Neural Networks [49.838283141381716]
  This paper addresses the challenge of deploying salient object detection (SOD) on resource-constrained devices with real-time performance.<n>We propose an efficient network design that combines traditional wisdom on SOD and the representation power of modern CNNs.
  arXiv  Detail & Related papers  (2025-07-01T20:41:05Z)
• S3MOT: Monocular 3D Object Tracking with Selective State Space Model [3.5047603107971397]
  Multi-object tracking in 3D space is essential for advancing robotics and computer applications.<n>It remains a significant challenge in monocular setups due to the difficulty of mining 3D associations from 2D video streams.<n>We present three innovative techniques to enhance the fusion of heterogeneous cues for monocular 3D MOT.
  arXiv  Detail & Related papers  (2025-04-25T04:45:35Z)
• MinkUNeXt: Point Cloud-based Large-scale Place Recognition using 3D Sparse Convolutions [1.124958340749622]
  MinkUNeXt is an effective and efficient architecture for place-recognition from point clouds entirely based on the new 3D MinkNeXt Block. A thorough assessment of the proposal has been carried out using the Oxford RobotCar and the In-house datasets.
  arXiv  Detail & Related papers  (2024-03-12T12:25:54Z)
• Towards Compact 3D Representations via Point Feature Enhancement Masked Autoencoders [52.66195794216989]
  We propose Point Feature Enhancement Masked Autoencoders (Point-FEMAE) to learn compact 3D representations. Point-FEMAE consists of a global branch and a local branch to capture latent semantic features. Our method significantly improves the pre-training efficiency compared to cross-modal alternatives.
  arXiv  Detail & Related papers  (2023-12-17T14:17:05Z)
• PonderV2: Pave the Way for 3D Foundation Model with A Universal Pre-training Paradigm [111.16358607889609]
  We introduce a novel universal 3D pre-training framework designed to facilitate the acquisition of efficient 3D representation.<n>For the first time, PonderV2 achieves state-of-the-art performance on 11 indoor and outdoor benchmarks, implying its effectiveness.
  arXiv  Detail & Related papers  (2023-10-12T17:59:57Z)
• GD-MAE: Generative Decoder for MAE Pre-training on LiDAR Point Clouds [72.60362979456035]
  Masked Autoencoders (MAE) are challenging to explore in large-scale 3D point clouds. We propose a textbfGenerative textbfDecoder for MAE (GD-MAE) to automatically merges the surrounding context. We demonstrate the efficacy of the proposed method on several large-scale benchmarks: KITTI, and ONCE.
  arXiv  Detail & Related papers  (2022-12-06T14:32:55Z)
• Towards Fast, Accurate and Stable 3D Dense Face Alignment [73.01620081047336]
  We propose a novel regression framework named 3DDFA-V2 which makes a balance among speed, accuracy and stability. We present a virtual synthesis method to transform one still image to a short-video which incorporates in-plane and out-of-plane face moving.
  arXiv  Detail & Related papers  (2020-09-21T15:37:37Z)
• A Real-time Action Representation with Temporal Encoding and Deep Compression [115.3739774920845]
  We propose a new real-time convolutional architecture, called Temporal Convolutional 3D Network (T-C3D), for action representation. T-C3D learns video action representations in a hierarchical multi-granularity manner while obtaining a high process speed. Our method achieves clear improvements on UCF101 action recognition benchmark against state-of-the-art real-time methods by 5.4% in terms of accuracy and 2 times faster in terms of inference speed with a less than 5MB storage model.
  arXiv  Detail & Related papers  (2020-06-17T06:30:43Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.