Leverage Cross-Attention for End-to-End Open-Vocabulary Panoptic Reconstruction

TL;DR

PanopticRecon++ introduces a novel end-to-end cross-attention approach with learnable 3D Gaussian queries for open-vocabulary 3D scene reconstruction, improving spatial priors and semantic consistency.

Contribution

It presents a new cross-attention formulation with learnable 3D Gaussian queries that enhance spatial priors and enable open-vocabulary panoptic reconstruction in an end-to-end manner.

Findings

Achieves competitive 3D and 2D segmentation and reconstruction results.

Effectively aligns 2D open-vocabulary instance IDs across frames.

Demonstrates applicability in robotic simulation scenarios.

Abstract

Open-vocabulary panoptic reconstruction offers comprehensive scene understanding, enabling advances in embodied robotics and photorealistic simulation. In this paper, we propose PanopticRecon++, an end-to-end method that formulates panoptic reconstruction through a novel cross-attention perspective. This perspective models the relationship between 3D instances (as queries) and the scene's 3D embedding field (as keys) through their attention map. Unlike existing methods that separate the optimization of queries and keys or overlook spatial proximity, PanopticRecon++ introduces learnable 3D Gaussians as instance queries. This formulation injects 3D spatial priors to preserve proximity while maintaining end-to-end optimizability. Moreover, this query formulation facilitates the alignment of 2D open-vocabulary instance IDs across frames by leveraging optimal linear assignment with instance masks rendered from the queries. Additionally, we ensure semantic-instance segmentation consistency by fusing query-based instance segmentation probabilities with semantic probabilities in a novel panoptic head supervised by a panoptic loss. During training, the number of instance query tokens dynamically adapts to match the number of objects. PanopticRecon++ shows competitive performance in terms of 3D and 2D segmentation and reconstruction performance on both simulation and real-world datasets, and demonstrates a user case as a robot simulator. Our project website is at: https://yuxuan1206.github.io/panopticrecon_pp/