Semantic-Guided Progressive Object Removal with Gaussian Splatting

2026-07-05

AI Summary

This work addresses the limitations of existing 3D Gaussian Splatting–based object removal methods, which struggle with complex geometry and textures and often fail to maintain cross-view consistency. To overcome these challenges, the paper introduces a novel framework that integrates multi-view semantic guidance with region-wise progressive refinement. Specifically, it first leverages DINOv2 to extract multi-view semantic features and employs a Semantic Block Matching (SBM) module to initialize missing regions. Subsequently, a Region-wise Progressive Refinement (RPR) strategy selectively reconstructs visually degraded subregions with enhanced detail. This approach significantly improves the fidelity of inpainted results in terms of geometric structure, texture richness, and multi-view consistency, outperforming current Gaussian-based methods in both perceptual quality and 3D coherence.

Abstract

Removing unwanted objects from reconstructed 3D scenes is an important task in computer vision, supporting applications in AR/VR, robotics, and digital content creation. Existing methods typically complete the entire masked region in a single step and without effectively utilizing semantic information from other views, leading to difficulties in handling complex geometric details and textures. In this work, we propose a novel framework that integrates Semantic-guided Block Matching (SBM) and Region-Wise Progressive Refinement (RPR) for high-quality 3D object removal. First, we leverage DINOv2 to encode semantic guidance from multi-view observations, and the best match tokens are decoded to complete missing regions in the target view while maintaining cross-view consistency. Second, we introduce a RPR strategy that segments the target mask into multiple subregions and selectively refines those with poor visual quality. Our method is built upon Gaussian Splatting, ensuring high-fidelity scene reconstruction with efficient computation. Experimental results demonstrate that our approach outperforms existing Gaussian-based methods in terms of perceptual quality and coherence in 3D object removal.

Problem

• 3D object removal
• semantic guidance
• Gaussian Splatting
• multi-view consistency
• scene completion

Innovation

• Semantic-guided Block Matching
• Region-Wise Progressive Refinement
• Gaussian Splatting
• Multi-view Consistency
• 3D Object Removal