arXiv 2606.089802026-06-08cs.CV

EPS3D: 端到端前馈式3D全景分割

EPS3D: End-to-End Feed-Forward 3D Panoptic Segmentation

Nanyang Technological University（南洋理工大学）

Runsong Zhu, Jiaxin Guo, Xiaoyang Guo, Zhengzhe Liu, Ka-Hei Hui, Wei Yin, Kai Chen, Wei Chen, Weiqiang Ren, Yunhui Liu, Pheng-Ann Heng, Chi-Wing Fu

备注：: ICML 2026. The code is publicly available at \href{https://github.com/Runsong123/EPS3D}{https://github.com/Runsong123/EPS3D}

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出端到端前馈框架EPS3D，通过蒸馏训练和多视图图像预测3D感知特征，结合互增强模块实现语义-实例一致性，在Replica上语义mIoU提升13%，每场景仅需1秒。

摘要：

本文介绍了EPS3D，一种用于开放词汇3D全景分割的新型端到端前馈框架。与依赖额外预处理的现有方法不同，我们设计了一种端到端架构，采用基于蒸馏的训练策略，在多样化的3D场景中从多视图图像预测3D感知的语义和实例特征，提高了3D一致性并避免了错误累积。我们进一步提出了一个互增强模块，以强制实现固有的语义-实例一致性。通过在实例内对齐语义（Ins2Sem）和利用语义指导细化实例特征（Sem2Ins），我们实现了更连贯的3D场景理解。最终，EPS3D在两个基准测试上优于最先进的基线（例如，在Replica上语义mIoU提升13%），且效率高（例如，每场景1秒），支持机器人操作和3D场景编辑等任务。

英文摘要：

This paper introduces EPS3D, a new end-to-end feed-forward framework for open-vocabulary 3D panoptic segmentation. Unlike existing methods relying on additional preprocessing, we design an end-to-end architecture, with a distillation-based training strategy on diverse 3D scenes to predict 3D-aware semantic and instance features from multi-view images, improving 3D consistency and avoiding error accumulation. We further propose a mutual enhancement module to enforce inherent semantic-instance consistency. By aligning semantics within instances (Ins2Sem) and refining instance features with semantic guidance (Sem2Ins), we achieve more coherent 3D scene understanding. Ultimately, EPS3D outperforms SOTA baselines on two benchmarks (e.g., +13% mIoU for semantics on Replica) with high efficiency (e.g., 1s per scene), supporting tasks like robotic manipulation and 3D scene editing.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。