Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

Announcing FPV-SLAM, a state-of-the-art SLAM system for egocentric data. FPV-SLAM ranks 1st in the monocular-inertial track of LaMAria, ETH Zurich's city-scale benchmark.

levelheaded_94's profile picture

Abhishek Anand

6,380 subscribers

13,647 Aufrufe • vor 12 Tagen •via X (Twitter)

Wissenschaft & Technologie
LIVE

Anya Rossi• Live Now

Private livecam show

30 Kommentare

Profilbild von Abhishek Anand
Abhishek Anandvor 12 Tagen

Read the full blog here:

Profilbild von Satpal Singh Rathore
Satpal Singh Rathorevor 12 Tagen

@fpv_labs Leading across all categories of short, med, long, low light, moving platform etc.

Profilbild von Gunjan Paul
Gunjan Paulvor 12 Tagen

@pratyush21225 @fpv_labs Congrats!!! No. 1 in LaMAria is crazy 🚀🚀

Profilbild von Abhishek Anand
Abhishek Anandvor 12 Tagen

@pratyush21225 @fpv_labs Thank you sir!

Profilbild von Asad Memon
Asad Memonvor 11 Tagen

@fpv_labs Congrats on the lead! Looking forward to reading about it (and improving Mighty Camera) 😬

Profilbild von Annu Shekhawat
Annu Shekhawatvor 12 Tagen

@fpv_labs Congratulations to the whole team! 🌟🔥

Profilbild von Abhishek Anand
Abhishek Anandvor 12 Tagen

@fpv_labs Thanks - we've come a long way since we published our first blog on the challenges with egocentric SLAM

Profilbild von Shubhanshu Khatana
Shubhanshu Khatanavor 12 Tagen

@fpv_labs LFG!!!🚀🚀🚀🚀

Profilbild von Abhishek Anand
Abhishek Anandvor 12 Tagen

@fpv_labs LFG!

Profilbild von ekaksh janweja
ekaksh janwejavor 12 Tagen

@fpv_labs congratulations 🔥

Profilbild von Abhishek Anand
Abhishek Anandvor 10 Tagen

@fpv_labs Thanks buddy!

Profilbild von Sourik
Sourikvor 12 Tagen

@fpv_labs Congratulations!

Profilbild von Abhishek Anand
Abhishek Anandvor 12 Tagen

@fpv_labs Thank you sir!

Profilbild von yogi
yogivor 12 Tagen

@fpv_labs congrats on the launch! lfgg

Profilbild von Abhishek Anand
Abhishek Anandvor 10 Tagen

@fpv_labs ♥️

Profilbild von mansin
mansinvor 12 Tagen

@fpv_labs Lfg!

Profilbild von Abhishek Anand
Abhishek Anandvor 10 Tagen

@fpv_labs LFG - we need to catchup soon!

Profilbild von Pratyush Patnaik
Pratyush Patnaikvor 12 Tagen

@fpv_labs lfg!! 🚀

Profilbild von Abhishek Anand
Abhishek Anandvor 10 Tagen

@fpv_labs LFGG!

Profilbild von Esteban Damian Fuhrmann
Esteban Damian Fuhrmannvor 11 Tagen

@fpv_labs Really impressive work! 🔥 Have you tested FPV-SLAM on long egocentric videos with lots of motion blur? I’d love to try it with our camera + IMU data!

Profilbild von Abhishek Anand
Abhishek Anandvor 10 Tagen

@fpv_labs sure can you pls email me at [email protected] - will connect with team who can help!

Profilbild von Anindyadeep
Anindyadeepvor 12 Tagen

@fpv_labs Lfgg congrats on the launch

Profilbild von Abhishek Anand
Abhishek Anandvor 12 Tagen

@fpv_labs Thanks buddy!

Profilbild von F-G Fernandez
F-G Fernandezvor 12 Tagen

@fpv_labs Congrats, mono+IMU on head-worn footage is the hard track: fast head turns, motion blur, people moving all around you. Is the ranked run online, or offline with full-sequence optimization? Both are useful, but they suit different products.

Profilbild von Srijan Jain
Srijan Jainvor 12 Tagen

@fpv_labs Congratulations!

Profilbild von Abhishek Anand
Abhishek Anandvor 10 Tagen

@fpv_labs Thanks sir!

Profilbild von vrushtee
vrushteevor 10 Tagen

@fpv_labs congratss!!

Profilbild von Abhishek Anand
Abhishek Anandvor 10 Tagen

@fpv_labs Thank youu!

Profilbild von Mahmood Ul Hassan, PhD
Mahmood Ul Hassan, PhDvor 10 Tagen

@fpv_labs City-scale monocular-inertial SLAM is a serious challenge! Impressive benchmark result for FPV-SLAM. Robust localization and loop closure in complex environments are key to scalable autonomous navigation.

Profilbild von Bruno Santos🇵🇹
Bruno Santos🇵🇹vor 10 Tagen

@fpv_labs Good job !

Ähnliche Videos

Today, we are announcing GroundedSLAM. LaMAria is the most difficult egocentric SLAM benchmark to date. We are the first to close the gap on Aria’s SLAM and eclipse all other methods by a wide margin. SLAM is the foundation of all 3D estimation. We think this is a big unlock for better robotics 👇🧵
0:35

Today, we are announcing GroundedSLAM. LaMAria is the most difficult egocentric SLAM benchmark to date. We are the first to close the gap on Aria’s SLAM and eclipse all other methods by a wide margin. SLAM is the foundation of all 3D estimation. We think this is a big unlock for better robotics 👇🧵

Grounded Superintelligence

101,416 Aufrufe • vor 1 Monat

Today we're announcing microSLAM, a monocular SLAM system built by our computer vision team in Zürich with ETH Zurich's Computer Vision and Geometry Lab. It ranks first on the LaMaria benchmark for monocular SLAM, and holds up against systems that carry many more cameras and IMUs. Ours runs on a single RGB stream. Robot learning is bottlenecked on data, and the largest untapped source is people going about their work. A camera on someone's head for an afternoon in a plant is a record of how that plant actually operates, including all the parts nobody writes down. That data is only useful if you can recover the geometry, and geometry from one moving camera in a world that will not sit still is the hard version of the problem. Solving it monocular is what makes the data cheap. Rigs do not scale to every worker. Glasses do. microSLAM turns that footage into interactive 3D environments where behavioral foundation models can be trained. One person walks a factory floor. A robot learns to navigate it.
3:30

Today we're announcing microSLAM, a monocular SLAM system built by our computer vision team in Zürich with ETH Zurich's Computer Vision and Geometry Lab. It ranks first on the LaMaria benchmark for monocular SLAM, and holds up against systems that carry many more cameras and IMUs. Ours runs on a single RGB stream. Robot learning is bottlenecked on data, and the largest untapped source is people going about their work. A camera on someone's head for an afternoon in a plant is a record of how that plant actually operates, including all the parts nobody writes down. That data is only useful if you can recover the geometry, and geometry from one moving camera in a world that will not sit still is the hard version of the problem. Solving it monocular is what makes the data cheap. Rigs do not scale to every worker. Glasses do. microSLAM turns that footage into interactive 3D environments where behavioral foundation models can be trained. One person walks a factory floor. A robot learns to navigate it.

Gianni Hodel

76,090 Aufrufe • vor 2 Monaten

All ready for the first Slam of the year! Rekortan 🤝 Grand Slam Track
0:47

All ready for the first Slam of the year! Rekortan 🤝 Grand Slam Track

Grand Slam Track

34,561 Aufrufe • vor 1 Jahr

This seemingly obvious prediction didn't take long to become reality. MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priors Contributions: • The first real-time SLAM system using the two-view 3D reconstruction prior MASt3R [20] as a foundation. • Efficient techniques for pointmap matching, tracking and local fusion, graph construction and loop closure, and second-order global optimization. • A state-of-the-art dense SLAM system capable of handling generic, time-varying camera models. Abstract: We present a real-time monocular dense SLAM system, designed from the ground up using MASt3R, a two-view 3D reconstruction and matching prior. Equipped with this strong prior, our system remains robust on in-the-wild video sequences, making no assumptions on a fixed or parametric camera model beyond a unique camera center. Key features include: - Efficient methods for pointmap matching, camera tracking, and local fusion - Graph construction and loop closure - Second-order global optimization With known calibration, a simple modification achieves state-of-the-art performance across various benchmarks. Altogether, we propose a plug-and-play monocular SLAM system capable of producing globally-consistent poses and dense geometry while operating at 15 FPS.
0:23

This seemingly obvious prediction didn't take long to become reality. MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priors Contributions: • The first real-time SLAM system using the two-view 3D reconstruction prior MASt3R [20] as a foundation. • Efficient techniques for pointmap matching, tracking and local fusion, graph construction and loop closure, and second-order global optimization. • A state-of-the-art dense SLAM system capable of handling generic, time-varying camera models. Abstract: We present a real-time monocular dense SLAM system, designed from the ground up using MASt3R, a two-view 3D reconstruction and matching prior. Equipped with this strong prior, our system remains robust on in-the-wild video sequences, making no assumptions on a fixed or parametric camera model beyond a unique camera center. Key features include: - Efficient methods for pointmap matching, camera tracking, and local fusion - Graph construction and loop closure - Second-order global optimization With known calibration, a simple modification achieves state-of-the-art performance across various benchmarks. Altogether, we propose a plug-and-play monocular SLAM system capable of producing globally-consistent poses and dense geometry while operating at 15 FPS.

MrNeRF

29,974 Aufrufe • vor 1 Jahr

Michael Johnson unveils new Track League to be called Grand Slam Track! - $12.6 million prize money in total - $100k for the winner of each slam - 4 Annual slams - Each slam is 3 days long - 96 athletes - No field events - No pacemakers
0:40

Michael Johnson unveils new Track League to be called Grand Slam Track! - $12.6 million prize money in total - $100k for the winner of each slam - 4 Annual slams - Each slam is 3 days long - 96 athletes - No field events - No pacemakers

Track & Field Gazette

172,346 Aufrufe • vor 2 Jahren

We are publishing our first deep dive on what we believe is one of the most challenging layers in egocentric data - SLAM and VIO in the context of long-horizon state tracking. We break down how SLAM and VIO fail in egocentric settings - visual features vanish at close range, depth sensors saturate, fast head motion blurs frames, and these failures don't always occur in isolation. They hit at the exact same moment, leading to compounding errors and making the downstream data unusable. We believe the foundation for high-quality egocentric data demands sub-centimeter precision over long episodes ranging from a few minutes to up to an hour.
0:33

We are publishing our first deep dive on what we believe is one of the most challenging layers in egocentric data - SLAM and VIO in the context of long-horizon state tracking. We break down how SLAM and VIO fail in egocentric settings - visual features vanish at close range, depth sensors saturate, fast head motion blurs frames, and these failures don't always occur in isolation. They hit at the exact same moment, leading to compounding errors and making the downstream data unusable. We believe the foundation for high-quality egocentric data demands sub-centimeter precision over long episodes ranging from a few minutes to up to an hour.

FPV Labs

42,211 Aufrufe • vor 7 Monaten

GIANNIS SLAM. GIANNIS SLAM. GIANNIS SLAM. THREE DUNKS IN A ROW FOR THE GREEK FREAK 💥💥💥
0:35

GIANNIS SLAM. GIANNIS SLAM. GIANNIS SLAM. THREE DUNKS IN A ROW FOR THE GREEK FREAK 💥💥💥

NBA

331,501 Aufrufe • vor 10 Monaten

🏆 NSAA TOURNAMENT UPDATE 🚨🚨DAKOTA STATE 2-OUT GRAND SLAM FOR THE LEAD Trojans 3B Dawson Portner delivers a go-ahead grand slam in the top of the 9th to extend the season for #5 DSU in a 5-4 victory over #8 Valley City State City State! Sam Tyrpa 4.2 IP 0 ER 7 K in relief DSU will take on the loser of #3 Dickinson State/ #7 Mayville State tomorrow 🎥: North Star #NAIABall
0:36

🏆 NSAA TOURNAMENT UPDATE 🚨🚨DAKOTA STATE 2-OUT GRAND SLAM FOR THE LEAD Trojans 3B Dawson Portner delivers a go-ahead grand slam in the top of the 9th to extend the season for #5 DSU in a 5-4 victory over #8 Valley City State City State! Sam Tyrpa 4.2 IP 0 ER 7 K in relief DSU will take on the loser of #3 Dickinson State/ #7 Mayville State tomorrow 🎥: North Star #NAIABall

NAIA Ball

17,569 Aufrufe • vor 1 Jahr

The threat of the slam can be as strong as the slam itself
0:32

The threat of the slam can be as strong as the slam itself

BM | Ticklememour @ Sandia Showdown 2026!!!!!

72,930 Aufrufe • vor 1 Jahr

Tom Hamilton’s Call of Carlos Santana’s GRAND SLAM in the GRAND SLAM GIVEAWAY INNING #GuardsBall
1:09

Tom Hamilton’s Call of Carlos Santana’s GRAND SLAM in the GRAND SLAM GIVEAWAY INNING #GuardsBall

k

61,192 Aufrufe • vor 1 Jahr

Malcolm Moore GRAND SLAM 💥 The 2024 1st rounder is hitting .360 with a 1.027 OPS for Hub City Spartanburgers in May
0:28

Malcolm Moore GRAND SLAM 💥 The 2024 1st rounder is hitting .360 with a 1.027 OPS for Hub City Spartanburgers in May

Rangers Player Development

35,573 Aufrufe • vor 6 Monaten

𝑪𝑨𝑳𝑳𝑰𝑵𝑮 𝑨𝑳𝑳 𝑹𝑰𝑽𝑨𝑳𝑺. This is the future of track. This is Grand Slam Track. Get in the Race: #GrandSlamTrack #OnlyTheFastest
0:40

𝑪𝑨𝑳𝑳𝑰𝑵𝑮 𝑨𝑳𝑳 𝑹𝑰𝑽𝑨𝑳𝑺. This is the future of track. This is Grand Slam Track. Get in the Race: #GrandSlamTrack #OnlyTheFastest

Grand Slam Track

154,862 Aufrufe • vor 2 Jahren

Monocular visual slam with loop closures on a $10 chip. cc Akshat
0:46

Monocular visual slam with loop closures on a $10 chip. cc Akshat

mansin

17,622 Aufrufe • vor 1 Monat

SEE YOU SOON, PHILLY 👊 Grand Slam Track heads to the City of Brotherly Love from May 30-June 1, 2025.
0:40

SEE YOU SOON, PHILLY 👊 Grand Slam Track heads to the City of Brotherly Love from May 30-June 1, 2025.

Grand Slam Track

16,642 Aufrufe • vor 2 Jahren