AI Insight
OmniAct3D is a new framework that adapts Vision Foundation Models for 3D object detection in 360-degree panoramic images using equirectangular projection. The system introduces three key components: an adapter for spherical geometry modeling, a reasoning chain that grounds detection hypotheses in panoramic evidence, and a heading estimation expert that processes object regions at higher resolution. Testing on benchmark datasets shows improvements of 2.96 NDS points on Spheriverse and 24.87 mAP points on PanoMMOcc compared to baseline methods, while maintaining 95-98% accuracy when adapted across different sensor configurations.
Why it matters
This research advances 3D perception capabilities for autonomous robots and vehicles that need comprehensive surround awareness. The framework's ability to leverage existing vision models for panoramic detection could accelerate development of mobile embodied AI agents while reducing the need for task-specific training data.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
Abstract: Accurate 3D detection is essential for mobile embodied agents, while Vision Foundation Models (VFMs) offer transferable visual and geometric priors. Yet existing VFM-based 3D detectors rely on narrow-view monocular images or discrete perspective views, limiting coherent surround perception; equirectangular projection (ERP) instead encodes a continuous 360 scene in a single image. Direct transfer remains difficult because ERP organizes geometry and visual information differently, making object-relevant cues hard to model, localize, and preserve. We propose OmniAct3D, a framework that adapts perspective-trained VFM detectors to ERP while preserving transferable VFM priors. To resolve geometric mismatch, the ERP-Ray Geometry Adapter (ERGA-Ray) models spherical viewing rays and periodic spatial structure. To localize evidence in scene-wide context, the Visual-Action Reasoning Chain (VARC) grounds each hypothesis in relevant panoramic evidence and converts it into a structured geometric action. To recover local cues lost under fixed token budgets, the Appearance-Guided Heading Expert (AGHE) re-encodes object regions at higher resolution for heading estimation. Experiments show that OmniAct3D improves over the previous best 3D detector by 2.96 NDS points on Spheriverse and over the unadapted VFM baseline by 24.87 mAP points on PanoMMOcc. With target-specific geometry adaptation, VARC retains 95–98% of the same-configuration mAP, indicating reusable object-level 3D reasoning across sensing configurations. The source code will be made publicly available at https://github.com/FeiT-FeiTeng/OmniAct3D.
Source: OmniAct3D: Leveraging Foundation Geometry and Evidence-Grounded Reasoning for Panoramic 3D Detection