HRoPE: Homography Relative Positional Encoding

Hugging Face Daily Papers Papers
positional-encoding multi-view-transformers novel-view-synthesis computer-vision homography geometry

Summary

HRoPE introduces a surface-aware positional encoding that models each patch as a slanted 3D plane and injects plane-induced homographies as relative transformations into attention, improving cross-view alignment and novel view synthesis without ground-truth geometry. Experiments on RealEstate10K, CO3Dv2, and DL3DV show it outperforms existing geometric positional encodings.

Geometric Relative Positional Encoding has become the primary mechanism for injecting spatial inductive biases into multi-view transformers. Existing designs condition attention on camera-level frustums or ray-marched depth, associating cross-view tokens through the sensor geometry. Such sensor-centric formulations, however, describe only where a token is observed from, leaving the rich surface geometry of the scene region depicted by the token. In this paper, we advocate a shift toward surface-aware positional encoding that explicitly grounds patch tokens in the geometry of physical surfaces. We introduce HRoPE, which represents the local geometry of each patch as a slanted 3D plane, following the slanted support-window assumption of classical multi-view stereo. We inject token-level, plane-induced homographies into attention as relative transformations. This equips the transformer with a geometric prior for rectifying perspective distortions and aligning patch features across viewpoints. Our encoding requires no ground-truth geometry: the plane parameters can be predicted end-to-end under a view-synthesis objective alone, or derived from foundation models. Extensive experiments on RealEstate10K, CO3Dv2, and DL3DV show that our best variant of HRoPE consistently outperforms the evaluated geometric positional encodings for novel view synthesis. Extensive ablations and analysis show the effectiveness of the proposed encoding. Please refer to our project page at: https://gaohchen.github.io/HRoPE
Original Article
View Cached Full Text

Cached at: 10/21/26, 02:21 PM

Paper page - HRoPE: Homography Relative Positional Encoding

Source: https://huggingface.co/papers/2610.28284

Abstract

HRoPE introduces surface-aware positional encoding using slanted 3D planes and homographies to improve cross-view alignment and novel view synthesis without requiring ground-truth geometry.

Geometric Relative Positional Encodinghas become the primary mechanism for injecting spatial inductive biases intomulti-view transformers. Existing designs condition attention on camera-level frustums or ray-marched depth, associatingcross-view tokensthrough the sensor geometry. Such sensor-centric formulations, however, describe only where a token is observed from, leaving the rich surface geometry of the scene region depicted by the token. In this paper, we advocate a shift toward surface-aware positional encoding that explicitly grounds patch tokens in the geometry of physical surfaces. We introduceHRoPE, which represents the local geometry of each patch as aslanted 3D plane, following the slanted support-window assumption of classical multi-view stereo. We inject token-level,plane-induced homographiesinto attention asrelative transformations. This equips the transformer with a geometric prior for rectifying perspective distortions and aligning patch features across viewpoints. Our encoding requires no ground-truth geometry: the plane parameters can be predicted end-to-end under a view-synthesis objective alone, or derived from foundation models. Extensive experiments on RealEstate10K, CO3Dv2, and DL3DV show that our best variant ofHRoPEconsistently outperforms the evaluated geometric positional encodings fornovel view synthesis. Extensive ablations and analysis show the effectiveness of the proposed encoding. Please refer to our project page at: https://gaohchen.github.io/HRoPE

View arXiv pageView PDFProject pageGitHub2Add to collection

Get this paper in your agent:

hf papers read 2610\.28284

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.28284 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.28284 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.28284 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

RIG-RoPE: Relation- and Instance-Gated Rotary Positional Encoding with Duration-Aware Temporal Coordinates

arXiv cs.CL

This preliminary technical report proposes RIG-RoPE, a relation- and instance-gated rotary positional encoding with duration-aware temporal coordinates, aiming to address spatial interference and improper temporal scaling in multimodal LLMs. It introduces a gating mechanism for height/width rotations and duration-aware temporal coordinates, but leaves large-scale empirical validation to future work.

RoVE: Rotary Value Embeddings Attention for Relative Position-dependent Value Pathways

arXiv cs.LG

This paper proposes RoVE, a parameter-free modification to Rotary Position Embeddings that makes value pathways position-sensitive by rotating values simultaneously with keys, transforming RoPE attention into attentive convolution. Experiments on GPT-2 models show consistent gains in few-shot in-context learning, out-of-distribution perplexity, and long-context retrieval.