UfM*: Uncertainty from Motion* for DNN Depth Estimation Using Gaussians
This paper proposes UfM*, an efficient uncertainty estimation algorithm that uses a compact Gaussian mixture to measure multiview disagreement from motion, requiring only a single DNN inference per image. It reduces calibration error by 24-28% compared to ensembles while consuming only 3% energy and 0.02% memory, enabling real-time operation on resource-constrained robots.
Article intelligence
Key points
- UfM* leverages motion to compute multiview disagreement via a Gaussian mixture, avoiding multiple inferences.
- Gaussian representation is more efficient and effective than point cloud for modeling 3D space disagreement.
- Combined with aleatoric uncertainty, UfM* improves calibration error by 24-28% over ensembles with minimal energy and memory.
- Running at 30 FPS on a low-power CPU, UfM* consumes only 63 mJ per frame, suitable for tiny robots.
Why it matters
This matters because ufM* leverages motion to compute multiview disagreement via a Gaussian mixture, avoiding multiple inferences.
Technical impact
May affect model selection, inference cost, product capability, and evaluation benchmarks.
[2605.23098] UfM*: Uncertainty from Motion* for DNN Depth Estimation Using Gaussians
[Submitted on 21 May 2026]
Title:UfM*: Uncertainty from Motion* for DNN Depth Estimation Using Gaussians
View a PDF of the paper titled UfM*: Uncertainty from Motion* for DNN Depth Estimation Using Gaussians, by Soumya Sudhakar and 2 other authors
View PDF HTML (experimental)
Abstract:Reliable uncertainty estimation is critical for deploying monocular depth deep neural networks (DNNs) in safety-critical robotic systems. Conventional uncertainty methods such as ensembles and sampling-based approaches require multiple inferences per image, incurring substantial compute and memory overhead. Moreover, uncertainty predicted from a single image misses out on measuring disagreement between predictions across views of the same region. We propose Uncertainty from Motion* (UfM*), an uncertainty estimation algorithm that measures multiview disagreement efficiently by comparing previous and current views using a compact Gaussian mixture, requiring only a single DNN inference per image. Using Gaussians to compute multiview disagreement is not only more compute- and memory-efficient than a prior approach using a point cloud, but also improves uncertainty by measuring disagreement across regions of 3D space. UfM* paired with aleatoric uncertainty improves expected calibration error by 24-28% compared to an ensemble, while requiring only 3% of the energy and 0.02% of the memory on 100 out-of-distribution ScanNet sequences. We demonstrate UfM* consumes only 63 mJ per 224x224 image while running real-time at 30 FPS on an Arm Cortex-A76 CPU onboard a miniature energy-constrained robot, highlighting that measuring multiview disagreement using Gaussians enables efficient uncertainty for resource-constrained robotic systems.
Comments: 18 pages, 15 figures
Subjects:
Robotics (cs.RO)
Cite as: arXiv:2605.23098 [cs.RO]
(or arXiv:2605.23098v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2605.23098
arXiv-issued DOI via DataCite (pending registration)
Submission history
From: Soumya Sudhakar [view email] [v1] Thu, 21 May 2026 23:08:42 UTC (7,492 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled UfM*: Uncertainty from Motion* for DNN Depth Estimation Using Gaussians, by Soumya Sudhakar and 2 other authors
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.RO
new | recent | 2026-05
Change to browse by:
cs
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
Loading...
Data provided by:
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)