OpenSU3D: Open World 3D Scene Understanding using Foundation Models
This addresses the scalability and generalization limitations in 3D scene understanding for applications like robotics and augmented reality, though it builds incrementally on existing foundation model techniques.
The paper tackles the problem of open world 3D scene understanding by developing a scalable approach that constructs instance-level 3D scene representations using 2D foundation models, achieving zero-shot generalization and exceeding state-of-the-art methods on ScanNet and Replica datasets.
In this paper, we present a novel, scalable approach for constructing open set, instance-level 3D scene representations, advancing open world understanding of 3D environments. Existing methods require pre-constructed 3D scenes and face scalability issues due to per-point feature vector learning, limiting their efficacy with complex queries. Our method overcomes these limitations by incrementally building instance-level 3D scene representations using 2D foundation models, efficiently aggregating instance-level details such as masks, feature vectors, names, and captions. We introduce fusion schemes for feature vectors to enhance their contextual knowledge and performance on complex queries. Additionally, we explore large language models for robust automatic annotation and spatial reasoning tasks. We evaluate our proposed approach on multiple scenes from ScanNet and Replica datasets demonstrating zero-shot generalization capabilities, exceeding current state-of-the-art methods in open world 3D scene understanding.