Pose-RFT: Enhancing MLLMs for 3D Pose Generation via Hybrid Action Reinforcement Fine-Tuning

• URL: http://arxiv.org/abs/2508.07804v1
• Date: Mon, 11 Aug 2025 09:44:58 GMT
• Title: Pose-RFT: Enhancing MLLMs for 3D Pose Generation via Hybrid Action Reinforcement Fine-Tuning
• Authors: Bao Li, Xiaomei Zhang, Miao Xu, Zhaoxin Fan, Xiangyu Zhu, Zhen Lei,
• Abstract summary: Pose-RFT is a reinforcement fine-tuning framework tailored for 3D human pose generation in MLLMs.<n>We introduce HyGRPO, a hybrid reinforcement learning algorithm that performs group-wise reward normalization over sampled responses.<n>Experiments on multiple pose generation benchmarks demonstrate that Pose-RFT significantly improves performance over existing pose-specific MLLMs.
• Score: 36.35364256178837
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Generating 3D human poses from multimodal inputs such as images or text requires models to capture both rich spatial and semantic correspondences. While pose-specific multimodal large language models (MLLMs) have shown promise in this task, they are typically trained with supervised objectives such as SMPL parameter regression or token-level prediction, which struggle to model the inherent ambiguity and achieve task-specific alignment required for accurate 3D pose generation. To address these limitations, we propose Pose-RFT, a reinforcement fine-tuning framework tailored for 3D human pose generation in MLLMs. We formulate the task as a hybrid action reinforcement learning problem that jointly optimizes discrete language prediction and continuous pose generation. To this end, we introduce HyGRPO, a hybrid reinforcement learning algorithm that performs group-wise reward normalization over sampled responses to guide joint optimization of discrete and continuous actions. Pose-RFT further incorporates task-specific reward functions to guide optimization towards spatial alignment in image-to-pose generation and semantic consistency in text-to-pose generation. Extensive experiments on multiple pose generation benchmarks demonstrate that Pose-RFT significantly improves performance over existing pose-specific MLLMs, validating the effectiveness of hybrid action reinforcement fine-tuning for 3D pose generation.

Related papers

• TIGaussian: Disentangle Gaussians for Spatial-Awared Text-Image-3D Alignment [58.46706158310462]
  TIGaussian harnesses 3D Gaussian Splatting (3DGS) characteristics to strengthen cross-modality alignment.<n>Our multi-branch 3DGS tokenizer decouples the intrinsic properties of 3DGS structures into compact latent representations.<n>A text-3D projection module adaptively maps 3D features to text embedding space for better text-3D alignment.
  arXiv  Detail & Related papers  (2026-01-27T06:30:32Z)
• MoRE: 3D Visual Geometry Reconstruction Meets Mixture-of-Experts [50.37005070020306]
  MoRE is a dense 3D visual foundation model based on a Mixture-of-Experts (MoE) architecture.<n>MoRE incorporates a confidence-based depth refinement module that stabilizes and refines geometric estimation.<n>It integrates dense semantic features with globally aligned 3D backbone representations for high-fidelity surface normal prediction.
  arXiv  Detail & Related papers  (2025-10-31T06:54:27Z)
• Gesplat: Robust Pose-Free 3D Reconstruction via Geometry-Guided Gaussian Splatting [21.952325954391508]
  We introduce Gesplat, a 3DGS-based framework that enables robust novel view synthesis and geometrically consistent reconstruction from unposed sparse images.<n>Our approach achieves more robust performance on both forward-facing and large-scale complex datasets compared to other pose-free methods.
  arXiv  Detail & Related papers  (2025-10-11T08:13:46Z)
• SPFSplatV2: Efficient Self-Supervised Pose-Free 3D Gaussian Splatting from Sparse Views [18.814209805277503]
  SPFSplatV2, an efficient feed-forward framework for 3D Gaussian splatting from sparse multi-view images, is presented.<n>Method achieves state-of-the-art performance in both in-domain and out-of-domain novel view synthesis.
  arXiv  Detail & Related papers  (2025-09-21T21:37:56Z)
• SeqAffordSplat: Scene-level Sequential Affordance Reasoning on 3D Gaussian Splatting [85.87902260102652]
  We introduce the novel task of Sequential 3D Gaussian Affordance Reasoning.<n>We then propose SeqSplatNet, an end-to-end framework that directly maps an instruction to a sequence of 3D affordance masks.<n>Our method sets a new state-of-the-art on our challenging benchmark, effectively advancing affordance reasoning from single-step interactions to complex, sequential tasks at the scene level.
  arXiv  Detail & Related papers  (2025-07-31T17:56:55Z)
• CoherenDream: Boosting Holistic Text Coherence in 3D Generation via Multimodal Large Language Models Feedback [40.163073128022944]
  Textual Coherent Score Distillation (TCSD) integrates alignment feedback from multimodal large language models (MLLMs)<n>3DLLaVA-CRITIC is a fine-tuned MLLM specialized for evaluating multiview text alignment in 3D generations.<n>CoherenDream establishes state-of-the-art performance in text-aligned 3D generation across multiple benchmarks.
  arXiv  Detail & Related papers  (2025-04-28T14:50:45Z)
• SIGMAN:Scaling 3D Human Gaussian Generation with Millions of Assets [72.26350984924129]
  We propose a latent space generation paradigm for 3D human digitization.<n>We transform the ill-posed low-to-high-dimensional mapping problem into a learnable distribution shift.<n>We employ the multi-view optimization approach combined with synthetic data to construct the HGS-1M dataset.
  arXiv  Detail & Related papers  (2025-04-09T15:38:18Z)
• StdGEN: Semantic-Decomposed 3D Character Generation from Single Images [28.302030751098354]
  StdGEN is an innovative pipeline for generating semantically high-quality 3D characters from single images.<n>It generates intricately detailed 3D characters with separated semantic components such as the body, clothes, and hair, in three minutes.<n>StdGEN offers ready-to-use semantic-decomposed 3D characters and enables flexible customization for a wide range of applications.
  arXiv  Detail & Related papers  (2024-11-08T17:54:18Z)
• AdaptivePose++: A Powerful Single-Stage Network for Multi-Person Pose Regression [66.39539141222524]
  We propose to represent the human parts as adaptive points and introduce a fine-grained body representation method. With the proposed body representation, we deliver a compact single-stage multi-person pose regression network, termed as AdaptivePose. We employ AdaptivePose for both 2D/3D multi-person pose estimation tasks to verify the effectiveness of AdaptivePose.
  arXiv  Detail & Related papers  (2022-10-08T12:54:20Z)
• Distribution-Aware Single-Stage Models for Multi-Person 3D Pose Estimation [29.430404703883084]
  We present a novel Distribution-Aware Single-stage (DAS) model for tackling the challenging multi-person 3D pose estimation problem. The proposed DAS model simultaneously localizes person positions and their corresponding body joints in the 3D camera space in a one-pass manner. Comprehensive experiments on benchmarks CMU Panoptic and MuPoTS-3D demonstrate the superior efficiency of the proposed DAS model.
  arXiv  Detail & Related papers  (2022-03-15T07:30:27Z)
• 3D Pose Estimation and Future Motion Prediction from 2D Images [26.28886209268217]
  This paper considers to jointly tackle the highly correlated tasks of estimating 3D human body poses and predicting future 3D motions from RGB image sequences. Based on Lie algebra pose representation, a novel self-projection mechanism is proposed that naturally preserves human motion kinematics.
  arXiv  Detail & Related papers  (2021-11-26T01:02:00Z)
• Direct Multi-view Multi-person 3D Pose Estimation [138.48139701871213]
  We present Multi-view Pose transformer (MvP) for estimating multi-person 3D poses from multi-view images. MvP directly regresses the multi-person 3D poses in a clean and efficient way, without relying on intermediate tasks. We show experimentally that our MvP model outperforms the state-of-the-art methods on several benchmarks while being much more efficient.
  arXiv  Detail & Related papers  (2021-11-07T13:09:20Z)
• PaMIR: Parametric Model-Conditioned Implicit Representation for Image-based Human Reconstruction [67.08350202974434]
  We propose Parametric Model-Conditioned Implicit Representation (PaMIR), which combines the parametric body model with the free-form deep implicit function. We show that our method achieves state-of-the-art performance for image-based 3D human reconstruction in the cases of challenging poses and clothing types.
  arXiv  Detail & Related papers  (2020-07-08T02:26:19Z)

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.