Pose-ICL: 3D-Aware In-Context Learning for Pose-Controllable Subject Customization

• URL: http://arxiv.org/abs/2606.10902v1
• Date: Tue, 09 Jun 2026 14:13:40 GMT
• Title: Pose-ICL: 3D-Aware In-Context Learning for Pose-Controllable Subject Customization
• Abstract summary: Pose-ICL is a tuning-free framework that adapts to new subjects through multiple paired image-pose references.<n>Its core mechanism,Surface-Anchored Position Embedding (SAPE), equips the model with explicit 3D awareness by anchoring image tokens to the surface coordinates of a volumetric bounding box.<n>Extensive evaluations on both 3D assets and real-world subjects demonstrate that Pose-ICL significantly outperforms current methods in both pose accuracy and identity consistency.
• Score: 9.596779036612409
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Subject Customization is a foundational task in modern image generation. By providing a few reference images and a text prompt, users can generate images of a specific object in any desired scene. However, existing methods still struggle to achieve effective pose control for customized subjects. In practice, they often exhibit inaccurate poses or inconsistent cross-pose appearances. These limitations suggest that understanding objects in a volumetric manner remains a significant challenge for 2D-native backbones. To address this challenge, we propose Pose-ICL, a tuning-free framework that leverages 3D-aware In-Context Learning (ICL) to directly adapt to new subjects through multiple paired image-pose references. Its core mechanism,Surface-Anchored Position Embedding (SAPE), equips the model with explicit 3D awareness by anchoring image tokens to the surface coordinates of a volumetric bounding box. Dedicated refinements ensure its seamless compatibility with existing DiT models. Extensive evaluations on both 3D assets and real-world subjects demonstrate that Pose-ICL significantly outperforms current methods in both pose accuracy and identity consistency.

Related papers

• Direct 3D-Aware Object Insertion via Decomposed Visual Proxies [101.76006356727015]
  DIRECT (Decomposed Injection for Reference Composition and Target-integration) is a novel framework that integrates interactive pose manipulation with high-fidelity 2D image synthesis.<n>Our method decomposes the insertion conditions into three complementary components: appearance guidance capturing visual details from the reference object, geometry guidance derived from the user-adjusted 3D proxy, and context guidance from the target background.<n>Experiments show that DIRECT outperforms previous methods in both geometric controllability and visual quality.
  arXiv  Detail & Related papers  (2026-06-04T18:00:34Z)
• Make-It-Poseable: Feed-forward Latent Posing Model for 3D Humanoid Character Animation [74.6792422278706]
  We introduce Make-It-Poseable, a novel feed-forward framework that reformulates character posing as a latent-space transformation problem.<n>Our method reconstructs the character in new poses by directly manipulating its latent representation.<n>It also naturally extends to 3D editing applications like part replacement and refinement.
  arXiv  Detail & Related papers  (2025-12-18T17:01:44Z)
• Articulate3D: Zero-Shot Text-Driven 3D Object Posing [38.75075284385844]
  We propose a training-free method, Articulate3D, to pose a 3D asset through language control.<n>We modify a powerful image-generator to create target images conditioned on the input image and a text instruction.<n>We then align the mesh to the target images through a multi-view pose optimisation step.
  arXiv  Detail & Related papers  (2025-08-26T17:59:17Z)
• PoseMaster: Generating 3D Characters in Arbitrary Poses from a Single Image [37.332231168919705]
  We propose PoseMaster, an end-to-end controllable 3D character generation framework.<n>Specifically, we unify pose transformation and 3D character generation into a flow-based 3D native generation framework.<n>Considering the specificity of multi-condition control, we randomly empty the pose condition and the image condition during training to improve the effectiveness and generalizability of pose control.
  arXiv  Detail & Related papers  (2025-06-26T08:03:14Z)
• ComposeAnything: Composite Object Priors for Text-to-Image Generation [72.98469853839246]
  ComposeAnything is a novel framework for improving compositional image generation without retraining existing T2I models.<n>Our approach first leverages the chain-of-thought reasoning abilities of LLMs to produce 2.5D semantic layouts from text.<n>Our model generates high-quality images with compositions that faithfully reflect the text.
  arXiv  Detail & Related papers  (2025-05-30T00:13:36Z)
• Visibility-Uncertainty-guided 3D Gaussian Inpainting via Scene Conceptional Learning [63.94919846010485]
  3D Gaussian inpainting (3DGI) is challenging in effectively leveraging complementary visual and semantic cues from multiple input views.<n>We propose a method that measures the visibility uncertainties of 3D points across different input views and uses them to guide 3DGI.<n>We build a novel 3DGI framework, VISTA, by integrating VISibility-uncerTainty-guided 3DGI with scene conceptuAl learning.
  arXiv  Detail & Related papers  (2025-04-23T06:21:11Z)
• 3D Congealing: 3D-Aware Image Alignment in the Wild [44.254247801001675]
  3D Congealing is a problem of 3D-aware alignment for 2D images capturing semantically similar objects. We introduce a general framework that tackles the task without assuming shape templates, poses, or any camera parameters. Our framework can be used for various tasks such as correspondence matching, pose estimation, and image editing.
  arXiv  Detail & Related papers  (2024-04-02T17:32:12Z)
• ObPose: Leveraging Pose for Object-Centric Scene Inference and Generation in 3D [21.700203922407496]
  ObPose is an unsupervised object-centric inference and generation model. It learns 3D-structured latent representations from RGB-D scenes. ObPose is evaluated quantitatively on the YCB, MultiShapeNet, and CLEVR datatasets.
  arXiv  Detail & Related papers  (2022-06-07T21:15:18Z)
• Self-Supervised Image Representation Learning with Geometric Set Consistency [50.12720780102395]
  We propose a method for self-supervised image representation learning under the guidance of 3D geometric consistency. Specifically, we introduce 3D geometric consistency into a contrastive learning framework to enforce the feature consistency within image views.
  arXiv  Detail & Related papers  (2022-03-29T08:57:33Z)
• Learning Pose-invariant 3D Object Reconstruction from Single-view Images [61.98279201609436]
  In this paper, we explore a more realistic setup of learning 3D shape from only single-view images. The major difficulty lies in insufficient constraints that can be provided by single view images. We propose an effective adversarial domain confusion method to learn pose-disentangled compact shape space.
  arXiv  Detail & Related papers  (2020-04-03T02:47:35Z)

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.