KiruPark

Robotics AI@Amazon. Robot Vision Ph.D at TU Wien. Former Ergonomic researcher@Hyundai Motors

[Anton Milan @antonmil]

I'm happy to share what we've been working on for the past few years.

[Amazon Science @AmazonScience]

Meet Sparrow. Amazon’s new intelligent robotic system that streamlines the fulfillment process by moving individual products before they get packaged, announced today at our #DeliveringTheFuture event: https://t.co/gbDj6fUi8o

[Srinath Sridhar @drsrinathsridha]

Neural fields are emerging as useful signal representations in computer vision & beyond. Our full-day introductory @CVPR tutorial on the topic is now public.

Video: https://t.co/e1EaySsOaI
Slides: https://t.co/AmTCAzIAO1
Web: https://t.co/zZAJwBNs7W

[Mohamed Sayed @MohammedAmr1]

📢 Our #ECCV2022 paper (and code) on fast accurate depth estimation and reconstruction is out now!

SimpleRecon: 3D Reconstruction without 3D Convolutions
https://t.co/dr0RjBBtIv

(1/4)

[Matthias Niessner @MattNiessner]

We released the videos & slides of our 3D Scanning & Motion Capture Lecture at TUM!

This was really fun to teach, covering fundamental 3D reconstruction and optimization techniques - enjoy watching :)

Videos: https://t.co/TgbZ8ATgq0
Slides & Content: https://t.co/8gWKlHYzGy

[KiruPark @Kirumang]

@StoiberMa Very impressive results and congratulations! If you don’t mind, would you tell me what camera you used for the demo in the video?

[KiruPark @Kirumang]

Definitely, ECCV deserves a lot more followers!

[European Conference on Computer Vision #ECCV2026 @eccvconf]

Only 4208 followers for one of #ComputerVision's top conferences!?!? We can do better. Please retweet

[KiruPark @Kirumang]

@quantombone It would be really fun if Infinite office provides a virtual group exercise at the beginning (or in the middle) of a long meeting!

[Amazon Science @AmazonScience]

Are you an MS or PhD student in Germany, interested in a 2022 applied science internship in the field of computer vision, or machine learning? Find out which teams at Amazon in Berlin are hiring interns for next year. #AmazonScience #Internships https://t.co/I44VYP6JhH

[Dave Clark @davehclark]

Also got to visit our robotic Air Hub sort center today which is very cool and the technical foundation for what we are installing in ground sort centers and the national air hub at CVG.

[Eric Brachmann @eric_brachmann]

Does the "ground truth" treat you fairly? Watch this shocking #ICCV21video 🤯🤯🤯 You will not believe what a 10-year-old baseline looks like today!

https://t.co/LbfHf4GUgA

#ICCV2021 w/ @martinhu @Carsten_Rother @SattlerTorsten

[AI at Meta @AIatMeta]

We train a network in an unsupervised manner that can, in a single feed forward pass, interpolate and establish dense correspondences between a pair of 3D shapes. https://t.co/tenDIfpYVU

[KiruPark @Kirumang]

@quantombone I was about to introduce the same paper above 😀

[Eric Brachmann @eric_brachmann]

RGB image in, set of 3D primitives out. A #CVPR2021 paper with @florian_kluger, H. Ackermann, M. Yang and B. Rosenhahn! #ComputerVision

abs: https://t.co/mVoz8VGtc4
code: https://t.co/Vhi7PixIbm

We take RANSAC out of its comfort zone into scene understanding territory. 👇

[Eric Brachmann @eric_brachmann]

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.

[Durk Kingma @dpkingma]

"The images are preprocessed to 256x256 resolution during training. [...] each image is compressed to a 32x32 grid of discrete latent codes using a discrete VAE that we pre-trained using a continuous relaxation."

GPT + VAE + scale = impressive results!

https://t.co/smbQcICNCL

[Justin Solomon @JustinMSolomon]

Want to learn about computer graphics? Feeling lonely during quarantine? In either case, check out my Intro to Computer Graphics course, where you can hear me talk about graphics nonstop for 26 hours, 33 minutes, and 11 seconds https://t.co/JuclUhZEfT (professionally captioned!)

[Ankur Handa @ankurhandos]

Generating synthetic scenes using Transformers. https://t.co/ciMeffp2Gq

Given an empty room, it figures out where to place an object (x, y, z, theta) and its size (l, w, h). All in an autoregressive manner (new object placement conditioned on the objects added already).

[KiruPark @Kirumang]

Strongly agree!! unless a paper proposes a method that needs a specific setup and metrics. This is what I usually have asked to papers that I’ve reviewed!

[Eric Brachmann @eric_brachmann]

#CVPR21 now, and I want to re-emphasise this 👇

If you work on 6DoF pose estimation, please please please use standardised evaluation procedures such as in the BOP challenge.

It is a MESS out there. Different metrics, diff. data splits, etc. all slapped together in one table.

[Krishna Murthy @_krishna_murthy]

Today we release gradslam - automagically differentiable SLAM. Run dense 3D reconstruction in @PyTorch!

And, it's fully differentiable :) Co-led w/ @S_Saryazdi, ably supported by @mautkiungli @duckietown_coo

Webage: https://t.co/vxVDedldRY
Paper: https://t.co/cN3tjc9e11