Saved in:
Bibliographic Details
Main Authors: Deng, Hui, Mao, Yuxin, He, Yuxin, Dai, Yuchao
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.14765
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908887925391360
author Deng, Hui
Mao, Yuxin
He, Yuxin
Dai, Yuchao
author_facet Deng, Hui
Mao, Yuxin
He, Yuxin
Dai, Yuchao
contents Streaming 3D reconstruction demands long-horizon state updates under strict latency constraints, yet stateful recurrent models often suffer from geometric drift as errors accumulate over time. We revisit this problem from a Grassmannian manifold perspective: the latent persistent state can be viewed as a subspace representation, i.e., a point evolving on a Grassmannian manifold, where temporal coherence implies the state trajectory should remain on (or near) this manifold.Based on this view, we propose Self-expressive Sequence Regularization (SSR), a plug-and-play, training-free operator that enforces Grassmannian sequence regularity during inference.Given a window of historical states, SSR computes an analytical affinity matrix via the self-expressive property and uses it to regularize the current update, effectively pulling noisy predictions back toward the manifold-consistent trajectory with minimal overhead. Experiments on long-sequence benchmarks demonstrate that SSR consistently reduces drift and improves reconstruction quality across multiple streaming 3D reconstruction tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2603_14765
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SSR: A Training-Free Approach for Streaming 3D Reconstruction
Deng, Hui
Mao, Yuxin
He, Yuxin
Dai, Yuchao
Computer Vision and Pattern Recognition
Streaming 3D reconstruction demands long-horizon state updates under strict latency constraints, yet stateful recurrent models often suffer from geometric drift as errors accumulate over time. We revisit this problem from a Grassmannian manifold perspective: the latent persistent state can be viewed as a subspace representation, i.e., a point evolving on a Grassmannian manifold, where temporal coherence implies the state trajectory should remain on (or near) this manifold.Based on this view, we propose Self-expressive Sequence Regularization (SSR), a plug-and-play, training-free operator that enforces Grassmannian sequence regularity during inference.Given a window of historical states, SSR computes an analytical affinity matrix via the self-expressive property and uses it to regularize the current update, effectively pulling noisy predictions back toward the manifold-consistent trajectory with minimal overhead. Experiments on long-sequence benchmarks demonstrate that SSR consistently reduces drift and improves reconstruction quality across multiple streaming 3D reconstruction tasks.
title SSR: A Training-Free Approach for Streaming 3D Reconstruction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.14765