Discover and read the best of Twitter Threads about #ICCV2021

Most recents (9)

Scale-Net: Learning to Reduce Scale Differences for Large-Scale Invariant Image Matching

tl;dr: same idea as with other ScaleNet: estimate scale difference in 2 images-> resize them. Eval on their version of Scale IMC
1/2 (check 2nd)
Yujie Fu, Yihong Wu
In order to prove priority and originality of the idea w.r.t. @axelbarrosotw ScaleNet, authors included #ICCV2021 submission screenshot as a proof.
Weird, however that current paper has less authors.
Read 3 tweets
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 πŸ§΅πŸ‘‡
Giorgos Tolias (@giotolias) from the CTU in Prague talked about Instance Search for Place Recognition:
Yannis Avrithis (@1avr1e) from the Athena Research Center talked about Metric Learning: Knowledge Transfer, Data Augmentation, and Attention:
Read 6 tweets
#ICCV2021 Old School/New School Panel

Richard Szeliski slide subset
Jitendra Malik slides
Great job @SergeBelongie! I could have watched another hour of this panel.
Read 3 tweets
We are presenting 5 papers at #ICCV2021! Get a first glimpse here:… More detailed info πŸ‘‡
(Just) A Spoonful of Refinements Helps the Registration Error Go Down, oral presentation, session 5A! @sragostinho @AljosaOsep Wonderful collaboration with @alessiodelbue
MOTSynth: How Can Synthetic Data Help Pedestrian Detection and Tracking?
Session 8A/8B! Paper: Project page:…
Read 6 tweets
🚨 This week at #ICCV2021, check out "Pixel-Perfect Structure-from-Motion with Featuremetric Refinement" w/ @PhilippCSE V. Larsson @mapo1
➑️ oral & *best student paper award*
thread ⬇️
We improve the accuracy of Structure-from-Motion (= camera poses & 3D points), making COLMAP subpixel accurate & closer to Lidar performance. This enables:
➑️ accurate visual localization of new images
➑️ mapping with fewer images - critical for large-scale crowd-sourced AR 😎
We align dense deep features with featuremetric optimization in 2 steps
1️⃣ Before SfM: keypoint adjustment
2️⃣ After SfM: bundle adjustment
This works well with many dense CNNs (even VGG ImageNet) & sparse local features (SuperGlue!) and suitable for both global/incremental SfM
Read 5 tweets
Excited to introduce a new generation of privacy protecting smart cameras in our #ICCV2021 Oral paper by @CarlosH_93, @henarfu and myself πŸ‡ΊπŸ‡ΈπŸ‡¨πŸ‡΄. See 🧡 below for details!


Our cameras introduce optical distortions at acquisition time, capturing images that protect the identity of people in the scene while enabling vision algorithms to perform inference accurately.

*Key idea*: joint end-to-end optimization of optical distortions & vision algorithm.
We achieve this by back-propagating all the way down to the camera lens. To make the lens shape differentiable, we parametrize it using Zernike polynomials.
Read 6 tweets
In anticipation of the Intl. Conf. on Computer Vision (#ICCV2021) this week, I rounded up all papers that use Neural Radiance Fields (NeRFs) represented in the main #ICCV2021 conference here (1/N):
Many of the papers I discussed in my original blog-post on NerF ( made it into CVPR, but the sheer number of NeRF-style papers that appeared on Arxiv this year meant I could no longer keep up. 2/N
Conferences like #ICCV2021 (with CVPR, the top-tier Computer Vision conference) provide an (imperfect) filter, and I decided to read all the papers I could find in the ICCV main program. I share them with you below and in the archival blog post. Email me if any are missing! 3/N
Read 46 tweets
Wondering how to detect when your neural network is about to predict pure non-sense in a safety critical scenario?

We answer your questions in our #ICCV2021 @ICCV_2021 paper!

Thursday 1am (CET) or Friday 6pm (CET), Session 12, ID: 3734


Thread πŸ§΅πŸ‘‡
The problem with DNNs is they are trained on carefully curated datasets that are not representative of the diversity we find in the real world.
That's especially true for road datasets.
In the real world, we have to face "unknown unkowns", ie, unexpected objects with no label.
How to detect such situation?
We propose a combination of 2 principles that lead to very good results:
1_ Disentangle the task (classification, segmentation, ...) from the Out-of-distribution detection.
2_ Train the detector using generated adversarial samples as proxy for OoD.
Read 10 tweets
Self-supervised learning promises to leverage vast amounts of data, but existing methods remain slow and expensive. Introducing contrastive detection, a new objective that learns useful representations for many tasks, with up to 10x less computation: 1/
Contrastive detection amplifies the learning signal from each image by carving it into pieces and learning from each simultaneously. This works particularly well when transferring to challenging tasks like detection, segmentation, and depth estimation. 2/
These results raise the possibility of learning algorithms that are more widely accessible, requiring neither human annotations nor vast amounts of computation.

To be presented at #ICCV2021 by @olivierhenaff @skandakoppula @jalayrac @avdnoord @OriolVinyalsML & @joaocarreira 3/3
Read 3 tweets

Related hashtags

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!