Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Liao, Guibiao, Ren, Qian, Liao, Kaimin, Wang, Hua, Chen, Zhi, Wang, Luchao, Tang, Yaohua
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2603.17519
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912972567216128
author Liao, Guibiao
Ren, Qian
Liao, Kaimin
Wang, Hua
Chen, Zhi
Wang, Luchao
Tang, Yaohua
author_facet Liao, Guibiao
Ren, Qian
Liao, Kaimin
Wang, Hua
Chen, Zhi
Wang, Luchao
Tang, Yaohua
contents Semantic-aware 3D reconstruction from sparse, unposed images remains challenging for feed-forward 3D Gaussian Splatting (3DGS). Existing methods often predict an over-complete set of Gaussian primitives under sparse-view supervision, leading to unstable geometry and inferior depth quality. Meanwhile, they rely solely on 2D segmenter features for semantic lifting, which provides weak 3D-level and limited generalizable supervision, resulting in incomplete 3D semantics in novel scenes. To address these issues, we propose UniSem, a unified framework that jointly improves depth accuracy and semantic generalization via two key components. First, Error-aware Gaussian Dropout (EGD) performs error-guided capacity control by suppressing redundancy-prone Gaussians using rendering error cues, producing meaningful, geometrically stable Gaussian representations for improved depth estimation. Second, we introduce a Mix-training Curriculum (MTC) that progressively blends 2D segmenter-lifted semantics with the model's own emergent 3D semantic priors, implemented with object-level prototype alignment to enhance semantic coherence and completeness. Extensive experiments on ScanNet and Replica show that UniSem achieves superior performance in depth prediction and open-vocabulary 3D segmentation across varying numbers of input views. Notably, with 16-view inputs, UniSem reduces depth Rel by 15.2% and improves open-vocabulary segmentation mAcc by 3.7% over strong baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2603_17519
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle UniSem: Generalizable Semantic 3D Reconstruction from Sparse Unposed Images
Liao, Guibiao
Ren, Qian
Liao, Kaimin
Wang, Hua
Chen, Zhi
Wang, Luchao
Tang, Yaohua
Computer Vision and Pattern Recognition
Semantic-aware 3D reconstruction from sparse, unposed images remains challenging for feed-forward 3D Gaussian Splatting (3DGS). Existing methods often predict an over-complete set of Gaussian primitives under sparse-view supervision, leading to unstable geometry and inferior depth quality. Meanwhile, they rely solely on 2D segmenter features for semantic lifting, which provides weak 3D-level and limited generalizable supervision, resulting in incomplete 3D semantics in novel scenes. To address these issues, we propose UniSem, a unified framework that jointly improves depth accuracy and semantic generalization via two key components. First, Error-aware Gaussian Dropout (EGD) performs error-guided capacity control by suppressing redundancy-prone Gaussians using rendering error cues, producing meaningful, geometrically stable Gaussian representations for improved depth estimation. Second, we introduce a Mix-training Curriculum (MTC) that progressively blends 2D segmenter-lifted semantics with the model's own emergent 3D semantic priors, implemented with object-level prototype alignment to enhance semantic coherence and completeness. Extensive experiments on ScanNet and Replica show that UniSem achieves superior performance in depth prediction and open-vocabulary 3D segmentation across varying numbers of input views. Notably, with 16-view inputs, UniSem reduces depth Rel by 15.2% and improves open-vocabulary segmentation mAcc by 3.7% over strong baselines.
title UniSem: Generalizable Semantic 3D Reconstruction from Sparse Unposed Images
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.17519