Fast and Reliable Open-Vocabulary Object Search with Absence Detection from Robot Memory
Fast and Reliable Open-Vocabulary Object Search with Absence Detection from Robot Memory
This paper introduces a two-stage architecture for mobile manipulators to perform fast and reliable open-vocabulary object search with absence detection. The system decouples candidate retrieval from verified selection, enabling high recall and specificity while operating in real time on a physical robot with a single consumer GPU. Evaluated on ScanNet apartment scenes including absent objects, the approach matches or exceeds exhaustive search reliability with significantly fewer detector calls compared to existing text or visual embedding methods that rarely abstain on missing targets.
The authors present a two-stage architecture for open-vocabulary object search that decouples fast candidate retrieval from verified selection, achieving high recall and reliable absence detection on ScanNet apartment scenes with physical robot validation using a single consumer GPU in real time.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Dataset details reported by the source
- Modalities
- RGB observations with pose Open source S4
- Embodiments
- mobile manipulators Open source S2
- paper: not_reported
What remains unknown
Dataset licensing and commercial-use rights are not established here.
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
equiring an order of magnitude fewer detector calls. We reproduce ScanNet scores on a physical robot operating entirely on-board with a single consumer GPU running in real time.
Open source S5
rieval from verified selection, enabling both high recall and reliable absence detection from sequential RGB observations with pose. We evaluate eight retrieval-selection configurations plus an exhaustive search baseline across ScanNet apartment scenes, including significant numbers of absent objects. Adding a visual verification step to caption-based retrieval matches or exceeds exhaustive search reliability while r
Open source S4
arXiv:2611.07493v1 Announce Type: new Abstract: Mobile manipulators deployed in everyday human environments, such as homes for assistive care or flexible automation, often need to locate specific objects on demand from accumulated observations. Equally important, they must reliably determine when a target is absent before committing to navigation or interaction. Existing approaches demand either slow exhaustive re-pr
Open source S2
ocessing or expensive 3D representations, and are evaluated on recall and accuracy rather than specificity, leaving their ability to abstain on missing targets unmeasured. We find existing systems that search using text or visual embedding almost never abstain on absent objects achieving at most 9\% specificity. To support accurate search at speed, we present a two-stage architecture that decouples fast candidate ret
Open source S3
Implications for data suppliers
RoboSignal interpretation and collection questions, not statements of buyer demand.
- Confirm the required data type and collection setting with the buyer; this source does not establish a complete collection specification.
- Check the exact dataset version, license, modalities and task coverage against the intended use.
- Validate demand and acceptance criteria with a buyer before scaling. Publication, popularity and a research result do not establish a purchase commitment.
Source：arXiv Robotics — research abstracts · arxiv.org