📢A new learning-based approach to SfM:
#ACEZero
No img-to-img matching, optimises image-to-scene correspondences directly. Needs no pose priors. Works on unordered image sets. Efficiently handles thousands of images.
Paper:
Page:
✨ Presenting MicKey (
#CVPR2024
, Oral) ✨
We regress and match 3D camera coordinates rather then 2D key points, all in metric space. Gives you a scaled relative pose between two RGB images.
Paper:
Page:
A
#CVPR2023
Highlight✨:
♥️♠️ACE: Accelerated Coordinate Encoding♣️♦️
We learn implicit, neural maps in minutes that let us relocalize with SOTA accuracy.
📜:
💻:
📽️:
work by
@NianticLabs
Research
Code of
#ACEZero
is out.
A new approach to SfM. Learn the 3D scene without image-to-image matching. Naturally avoids the explosion of complexity for many images. ACE0 shines if you have dense coverage of a scene. Posing 10k images and more? Sure!
We've published
@pytorch
code of Differentiable RANSAC for a toy problem: fitting lines. A CNN learns to predict points (middle) to which we robustly fit lines, trained end2end with DSAC. Right: A CNN which learns to predict line parameters directly. Code:
Read papers to write good papers.
Review papers to write great papers.
Read rebuttals to write good rebuttals.
Discuss rebuttals to write great rebuttals.
Great papers and great rebuttals are no guarantee to get accepted. Iterations are normal.
Ten years ago, my first top-tier paper was accepted to
#ECCV2014
. It was about learning to predict dense correspondences for object pose. It spawned a line of work that lead to
#ACEZero
- there are still traces of the old code. Happy that ACE0 was accepted to
#ECCV2024
! 🔄
Today, we launch Map-Free Visual Relocalisation. A twist on the usual relocalisation formula. A new dataset. A new benchmark.
Check out our video!
#MapFreeReloc
to be presented at
#ECCV2022
Niantic organises the first
#MapFreeReloc
workshop and challenge at
#ECCV2024
@eccvconf
!
We invite the submission of extended abstracts and papers. The challenge features two tracks, and includes a multi-frame version and $6000 in prizes.
Page:
We created some visualisations of MASt3R on the
#MapFreeReloc
dataset. The numbers do not lie, the results are amazing.
The public version is a bit worse than the private one which is top on the leaderboard. 0.751 AUC instead of 0.817.
Examples in 🧵including public vs private.
Major update and code release for our
#relocalization
pipeline, DSAC*
(Updated) paper:
Code:
Use RGB or RGB-D, train with an SfM model or a 3D scan. Train from images and poses alone and DSAC* discovers coarse geometry by itself:
We dedicate our oral presentation at
#CVPR2024
in Seattle to R2, who - in a great moment of integrity and reason - raised their initial "weak reject" rating to "strong reject" after rebuttal.
We hope they enjoy every minute of our 15 minute talk.
✨ Presenting MicKey (
#CVPR2024
, Oral) ✨
We regress and match 3D camera coordinates rather then 2D key points, all in metric space. Gives you a scaled relative pose between two RGB images.
Paper:
Page:
RGB image in, set of 3D primitives out. A
#CVPR2021
paper with
@florian_kluger
, H. Ackermann, M. Yang and B. Rosenhahn!
#ComputerVision
abs:
code:
We take RANSAC out of its comfort zone into scene understanding territory. 👇
With the
#CVPR2023
content being public now, so are the recordings of the talks I gave.
Firstly, I gave a talk at the
#IMC2023
workshop:
"Pose Estimation Beyond Feature Matching"
I recommend watching the whole session of this awesome workshop!
1/2
I really like the Twitter format for presenting papers. So I thought I replicate that for my personal website (neglected for years...)
It's a simple static HTML, feel free to use as template.
Due to requests at
#ECCV2022
and to make our
#MapFreeReloc
dataset useful for more tasks, we make the SfM reconstructions of our train set publicly available.
🔥460 SfM models of outdoor scenes all around the world 🔥
Want to train 460 NeRFs? Go ahead.
All talks of our
#ICCV2021
tutorial on visual localisation are available on YouTube.
In my talk about learning-based localisation, I discuss a challenge, a promise and a compromise.
Watch it here:
Links to all the amazing talks in 🧵👇
🔥Map-relative Pose Regression🔥(
#CVPR2024
highlight)
For years absolute pose regression did not work. There was some success by massively synthesising scene-specific data. We train scene-agnostic APR and it works.
Paper:
Page:
Third and last code package for NG-RANSAC (
#ICCV2019
) is online: NG-DSAC++ for camera re-localization, a re-implementation of DSAC++ in
@PyTorch
, extended with neural guidance.
Code:
Paper:
w/
@Carsten_Rother
@LabHeidelberg
We are hiring! Want to work with me and an amazing team at
@NianticEng
on re-localisation at a global scale? Bring cutting-edge research into the hands of millions of users?
Consider applying for our Mapping and Localisation MLE role:
#ACEZero
is great to pose every frame of a video. But why not use SLAM? Sometimes it's advantageous to have no frame-to-frame assumption. No need to worry about loosing tracking.
In this example, we register "only" 70% of the frames. The rest is dismissed. Good or bad? Depends.
Niantic research has a strong presence at
#CVPR2023
with 5 papers (one highlight) and various contributions throughout workshops and tutorials.
Diffusion, NeRF, relocalisation, object pose, feature matching, depth and occlusions.
Here is where you can catch us:
We uploaded estimated poses of
#ACEZero
for the main paper experiments. If you do want to compare without running our code.
You also get the reconstruction videos, 78 in total.
Links in the README:
Eg reconstructing T&T Caterpillar from 11k images.
2 papers accepted to
#cvpr2020
🥳
Reinforced Feature Points: Use classic REINFORCE to optimize feature detection and description for the task you care about
CONSAC: fit multiple parametric models by learned sequential search (w/
@florian_kluger
)
More info soon!
#ComputerVision
#ACEZero
can be slower than other methods for sparse view reconstruction. That's a downside of the incremental learning approach. But methods can be combined.
Here we start from sparse poses of COLMAP, and then densify and refine with ACE0. ACE0 repairs drift along the way, too.
When you try to solve difficult image pairs, it's important that you do not overshoot and start to hallucinate connections between unrelated images. The
#MapFreeReloc
benchmark checks for that. The inlier count of MicKey is pretty good in separating solvable and unsolvable cases.
.
@SattlerTorsten
talking about "Old School" methods at the
#ICCV2021
tutorial on visual localization. "Old" but not outdated!
Join us here:
I will talk about the "New School" later.
Some motivation if you didn't make it into
#CVPR
:
- our first object coordinates paper: rejected from CVPR14, barely made it into ECCV14 (>250 citations)
- follow up: rejected from ICCV15, made it into CVPR16 (~200 citations)
- ESAC: rejected from CVPR19, made it into ICCV19
Catch Niantic Research at
#CVPR2024
!
- I will give a talk at the MonoDepth challenge, TUE-PM. Teaser below!
- Niantic presents two posters WED-PM: AirPlanes and MicKey. Do not miss the MicKey 📢Oral 📢 by
@axelbarrosotw
!
- One more poster FRI-AM: marepo (✨Highlight✨)
New retro wave of
#ComputerVision
: NG-RANSAC brings you the greatest hits of the 80s: RANSAC, multi-layer perceptrons, (classic) reinforcement learning. Lens flare for visualization of "cool", only. The paper:
#ICCV2019
#ICCV19
#ICCV
#DeepLearning
To
#CVPR2023
reviewers: Remember there are humans on the other end. Be strict in the matter but respectful in tone. Consider even being friendly in tone. Someone worked hard on this, and is proud. Don't bend but sweeten the pill.
With the
#ACErelocalizer
, we reduced mapping times from 15 hours to 5 minutes. You can imagine that further speed improvements are exponentially harder. Yet, we managed to squeeze out another 15% speed-up. Let me walk you through the steps:
1. Upgrade to
@pytorch
2.x
Fin.
A
#CVPR2023
Highlight✨:
♥️♠️ACE: Accelerated Coordinate Encoding♣️♦️
We learn implicit, neural maps in minutes that let us relocalize with SOTA accuracy.
📜:
💻:
📽️:
work by
@NianticLabs
Research
Automatic generation of ground truth is great but caution is advised.
Upcoming for
#ICCV2021
: For vis. relocalisation, we show that depending on how you generate GT, the ranking of relocalisers flips upside down:
@SattlerTorsten
@martinhu
@Carsten_Rother
Submitting relocalisation via pose regression to
#ICCV2023
? Rejected from
#CVPR2023
? Consider the
#MapFreeReloc
benchmark.
No need to beat DSAC*, hLoc, AS... they do not apply. This benchmark was created just for you 🫵
Also if you work on:
- features
- depth
- uncertainty
🧵
Unfortunately, none of the authors of this great paper was able to make it to
#CVPR2024
. I was able to put it up, but have to go to our other poster now. So let's give this poor poster some love on Twitter at least!
We uploaded alternative "ground truth" and full SfM models for the
#relocalisation
datasets 7Scenes and 12Scenes.
Working on reloc towards
#ECCV2022
and having trouble beating SOTA? The "ground truth" might play a role.
#ICCV2021
#betterlatethannever
Results of the current runner-up on the
#MapFreeReloc
leaderboard (single frame):
MicKey (
#CVPR2024
).
Code is available, might be a good starting point for the
#ECCV2024
map-free challenge!
When I first saw MASt3R on our
#MapFreeReloc
leaderboard, I thought someone was trolling us. The name did not help. After reading the paper and chatting to the authors, I believe it's legitimate.
What incredible progress on a new task in ~2 years. And more is possible, I'm sure!
Grounding Image Matching in 3D with MASt3R
@Vinc3nt_Leroy
Yohann Cabon
@JeromeRevaud
tl;dr: Dust3r with descriptor head and metric depth.
P.S. guys, why not some KD-tree, why new fast knn?
MAST3R leads the
#MapFreeReloc
leaderboard by a large margin. Glad that the code is out, so people can improve on it 💪 43 more days to compete in the
#MapFreeReloc
challenge for
#ECCV2024
!
The wait is over 📢 MAST3R is out! DUSt3R+ dense local feature maps & metric depth - 1st in
#MapFreeReloc
leaderboard, can handle 1000s of images 😀 !!
Blog:
Code:
Paper:
The moment has finally come. I tried to get the best education possible to prepare for this. Worked hard. Let's see whether it all payed off.
My son learned the word "why".
Personally, I find these results all the more remarkable considering that MicKey does not use any cross attention. Key points and descriptors are predicted per image, without considering the other view. Just vanilla matching and RANSAC after that.
Beyond providing metric estimates, MicKey can also deal with extreme view point changes, up to opposing shots. Here are a few examples of MicKey correspondences.
Stumbled across this
#ICCV2023
paper extending
#MapFreeReloc
to panoramic indoor views. Makes a lot of sense: A panorama serves as a one-shot map with wide coverage.
Calibrating Panoramic Depth Estimation for Practical Localization and Mapping, Kim et al
If you like these videos, we just pushed the code to generate them to the
#MapFreeReloc
repo. Might be handy for those who work towards the
#ECCV2024
challenge.
We created some visualisations of MASt3R on the
#MapFreeReloc
dataset. The numbers do not lie, the results are amazing.
The public version is a bit worse than the private one which is top on the leaderboard. 0.751 AUC instead of 0.817.
Examples in 🧵including public vs private.
Learning-based visual relocalisation has made quite some progress over the years.
I'm giving an overview in our
#CVPR2023
tutorial on:
📢"Large-Scale Visual Localisation"
Mark June 19th in your calendar. Check out the full schedule of amazing talks:
Strange. My Twitter feed somehow suppresses all the tweets of PhD students from small labs that complain about the
#CVPR2022
social media motion. I see, however, plenty tweets by 1k+ follower accounts complaining that the former group will be at a disadvantage now 🙃
It's fun playing around with
#ACEZero
. Just point it to a set of images, and you are good to go.
(The Lego model? Yeah, I built that. Don't mention it. The middle section? Yeah, it can totally detach a stand-alone Ninja bike! Pretty cool? Well, if you say so... )
Visual relocalization has been a bit of a niche topic of 3D vision. I hope this work can help it reaching a wider audience by showing how tightly coupled it is with reconstruction.
People think of reloc as a step after or within incremental SfM.
#ACEZero
turns that on its head.
📢A new learning-based approach to SfM:
#ACEZero
No img-to-img matching, optimises image-to-scene correspondences directly. Needs no pose priors. Works on unordered image sets. Efficiently handles thousands of images.
Paper:
Page:
You can use
#ACEZero
to refine a set of existing poses. In the example below, we start from KinectFusion poses. ACE0 repairs a defect in KF tracking, and the ACE0 refinement gains 4dB in PSNR for a Nerfacto model on top.
Everything tidied up in 8 minutes for 6000 frames 😇
What's
#MapFreeReloc
?
A learned model induces a scale-metric space, conditioned on a single reference frame. We localise new queries in that space.
I will talk about "Pose Estimation Beyond Feature Matching" on Monday morning at
#CVPR2023
at the Image Matching Workshop
#IMC2023
I understand why ICCV/ECCV rank lower than CVPR because they are not annual. But why does ECCV rank lower than ICCV?
It would appear to me they recruit chairs and reviewers from roughly the same pool. ECCV is the second important deadline in even years, as ICCV is for odd years.
Dear
#ECCV2024
reviewer,
"see weaknesses" is not a good "justification of rating". Every paper has weaknesses. The AC needs to understand why you think the weaknesses outweigh the strengths (or vice versa).
Best,
You would assume that since 2012
#DeepLearning
was applied to everything and their mother. But, as far as we know, CONSAC is the first learned multi-model robust estimator.
By
@florian_kluger
#CVPR2020
arXiv:
code:
Our call for research interns
@NianticEng
for 2022 is live! Interested in pushing the frontiers of AR? So are we. Work with us on a wide range of topics in the spectrum of
#ComputerVision
and
#MachineLearning
!
Interested? More info and application forms:
Today, Wednesday morning, at
#ICCV2019
, I will present NG-RANSAC at poster stand 143 (far end of the expo hall). I will try to explain the method on a simple toy problem, such that even I could understand it :) Come by, say hello!
#ICCV
#ICCV19
#DeepLearning
#MachineLearning
Two papers accepted to
@ICCV19
! Neural-Guided RANSAC (NG-RANSAC): A neural network guiding RANSAC data point selection, and Expert Sample Consensus (ESAC): An ensemble of scene coordinate experts for scalable camera re-localization.
#ICCV2019
#ComputerVision
#DeepLearning
I keep reading in papers that RANSAC is so excruciatingly slow that getting rid of it is a contribution.
Watch DSAC* estimating poses from thousands of correspondences in 30ms per frame, including network execution and RANSAC. 👇
Brrr.
Extending on my earlier post about research trends at
#CVPR2018
, I wrote a little python script that plots topic popularity (measured by key word matches against paper titles) over time. It's a simple Jupyter notebook, so you can play around yourself:
You know these annoying 16bit PNGs, eg depth maps, that your default image viewer probably does not show correctly? Did you know that
#ImageJ
can open them? Did you also know that
#ImageJ
has been ported to JavaScript and runs in your browser!? 😲
There is this magical moment of excitement and joy, right after drafting a new approach, and before actually implementing it and seeing it fail in all the experiments.
Thanks
@ducha_aiki
for sharing our work faster than we could get it to arXiv. Its available there as well, now.
📜
TL;DR? Upload the gist of it into your brain via YouTube:
▶️
w/
@axelbarrosotw
,
@viprad
, Gabe, and
@dantkz
Two-view Geometry Scoring Without Correspondences
Axel Barroso-Laguna,
@eric_brachmann
Victor Adrian Prisacariu Gabriel Brostow
@dantkz
tl;dr: in title + nice analysis of the metrics for the epipolar geometry.I almost want to write a blogpost-review :)
📢📢 Announcing the 8th workshop on Recovering 6D Object Pose (
#R6D
) at
#ICCV2023
📢📢 CFP: We invite paper submissions of unpublished works covering object pose estimation and related topics. Deadline: July 24
#BOP
challenge 2023 also in the works!
I really like that I'm not able to rate "borderline" in the final
#ICCV2021
assessment. I'm split regarding several submissions, now I have to think twice as I have to lean in some direction. An unexpectedly enjoyable inner fight.
6/6
#CVPR2022
reviews done. Strong stack this one: worst score is borderline - never had this before. Innovative ideas and above-average writing. I'm impressed.
Finally, there is an extensive benchmark for 6D object pose estimation, presented at
#ECCV2018
. No learned method in top 3. Best learned method uses Random Forests instead of CNNs. Want to defend the honor of deep learning? It's a running competition ;)
Naver Labs is joining the
#MapFreeReloc
challenge as co-sponsor, giving $2000!
Thanks
@JeromeRevaud
and
@naverlabseurope
for making it happen!
We have a total of $6000 in prizes for the challenge winners.
The slides for my talk at the R6D workshop at
#iccv2019
are available at the workshop website:
A summary of our work on differentiating PnP, differentiable RANSAC, differentiable correspondence selection and differentiable expert selection.
#DeepLearning
A small thread on 3D rotations: Both log-quaternions (log-q) and axis-angles (aa) represent rotations with 3 parameters. But they are not the same, related by a factor of 2. The length of aa gives you the rotation angle, the length of log-q gives you half that angle.
📢 Benchmark for 6D Object Pose Estimation 📢
#BOP
challenge 23 opened!
Results to be presented at the
#R6D
workshop at
#ICCV2023
.
The challenge has always pushed the field forward. This year: ❓unseen objects❓. Can you onboard a new object in 5mins?
This summer I've been working to finally understand Lie Theory, the basis for proper estimation on over-parameterised manifolds like SE(3). There are some great tutorials for the roboticist out there; I especially like Micro Lie Theory by Solà et al.
On Wednesday I will be giving a talk at
@naverlabseurope
on visual localization. I will talk about differentiable RANSAC and new stuff, including learning to guide RANSAC and scalable re-localization. If you are in Grenoble, come by!
To add to my list of motivating failures: I had 2 submissions lined up for
#ECCV2020
. One stopped a few weeks ago, one stopped a few hours before deadline. So that happens, too!
At the same time, this is my contribution to
#slowscience
. You are welcome 😜
Some motivation if you didn't make it into
#CVPR
:
- our first object coordinates paper: rejected from CVPR14, barely made it into ECCV14 (>250 citations)
- follow up: rejected from ICCV15, made it into CVPR16 (~200 citations)
- ESAC: rejected from CVPR19, made it into ICCV19
Like so many others, I also had a lot of fun at
#CVPR2024
! It was great to chat with old friends and meet new people (Yes, we will release the ACE0 code soon!) 5 days of conference are still not enough to catch up with everybody... Hope to make up for some of that in Milano.