Alex Dailo

XR @ Qualcomm

[Alex Dailo @ukrdailo]

MC-Calib: a generic and robust calibration toolbox for multi-camera systems

TLDR: open source to calibrate multi-camera (perspective, fisheye, hybrid) systems with diverse configurations (overlapping, non-overlapping, converging).

GitHub: https://t.co/Gz5uuLXhog

[Alex Dailo @ukrdailo]

If you're at #CVPR2026, come see Fast SceneScript in person tomorrow morning!

Learn how we hit 5× faster 3D scene layout estimation using multi-token prediction.

📍 Poster #227 (ExHall A-F)
⏰ 10:45 AM - 12:45 PM | Poster Session 1

[Alex Dailo @ukrdailo]

Excited to share that our paper “Fast SceneScript” is accepted to @CVPR 🎉

5× faster language-based 3D scene layout estimation via multi-token prediction — without sacrificing accuracy.

https://t.co/z6rUYXcB9g
@RuihongYin12 @XuepengShi123 @Marco__Manfredi

[Alex Dailo @ukrdailo]

@ducha_aiki @Christopher_Xie @yawarnihal @jstraub6 @rapideRobot Scenescript is extremely powerful as it can be extended to different representations only by expanding the language.

We've actually optimized it 5x without sacrificing accuracy with multi-token prediction, resulting in FastScenescript https://t.co/z6rUYXcB9g

[Alex Dailo @ukrdailo]

Just got to know about @gistdotscience and the amazing paper summary they did for Fast SceneScript https://t.co/WYHvR3ZcPf

[Alex Dailo @ukrdailo]

Thanks to @CVPR reviewers, we extended the paper with validation on real-world data and experiments on 3D object detections. The original findings remain consistent.

Paper: https://t.co/z6rUYXcB9g

[Alex Dailo @ukrdailo]

Excited to share that our paper “Fast SceneScript” is accepted to @CVPR 🎉

5× faster language-based 3D scene layout estimation via multi-token prediction — without sacrificing accuracy.

https://t.co/z6rUYXcB9g
@RuihongYin12 @XuepengShi123 @Marco__Manfredi

[Alex Dailo @ukrdailo]

Excited to share that our paper “Fast SceneScript” is accepted to @CVPR 🎉

5× faster language-based 3D scene layout estimation via multi-token prediction — without sacrificing accuracy.

https://t.co/z6rUYXcB9g
@RuihongYin12 @XuepengShi123 @Marco__Manfredi

[Alex Dailo @ukrdailo]

Exciting 6-month internship at @QCOMResearch #XR Amsterdam in Deep Learning for 3D Scene Understanding. Join a team pushing the boundaries of AI for 3D computer vision and extended reality. Learn more and apply here: https://t.co/VgHulhqO3j

[Alex Dailo @ukrdailo]

With the final release of 2025, MC-Calib v1.4.0 now also works with the latest @opencvlibrary v4.10 and comes with multiple other improvements (see https://t.co/20aExd2KLT)

[Alex Dailo @ukrdailo]

MC-Calib: a generic and robust calibration toolbox for multi-camera systems

TLDR: open source to calibrate multi-camera (perspective, fisheye, hybrid) systems with diverse configurations (overlapping, non-overlapping, converging).

GitHub: https://t.co/Gz5uuLXhog

[Naamii Nepal @naamii_nepal]

Meet @Frameau89 , Assistant Professor at SUNY in Korea, who was a speaker at ANAIS 2023. He will be joining us again this year at 5th ANAIS 2024, so don't miss your chance to learn from this incredible speaker, along with 14+ other speakers from around the world!

[Markus Nagel @mnagel87]

Are you pursuing a PhD and are you interested in working on efficiency of LLMs/LVMs? Then join our model efficiency team in #QualcommAIResearch for an internship!

Apply below, we have openings for 2025 as well as autumn/winter 2024.
https://t.co/vNrUT9wzrB

[Alex Dailo @ukrdailo]

MC-Calib v1.3.0 comes with improvements to the utilities for post-calibration analysis as well as bug fixes in non-overlapping camera calibration scenario

https://t.co/9HzxOnmn6h

[Alex Dailo @ukrdailo]

MC-Calib: a generic and robust calibration toolbox for multi-camera systems

TLDR: open source to calibrate multi-camera (perspective, fisheye, hybrid) systems with diverse configurations (overlapping, non-overlapping, converging).

GitHub: https://t.co/Gz5uuLXhog

[OpenAI @OpenAI]

Say hello to GPT-4o, our new flagship model which can reason across audio, vision, and text in real time: https://t.co/MYHZB79UqN

Text and image input rolling out today in API and ChatGPT with voice and video in the coming weeks.

[Alex Dailo @ukrdailo]

Cannot overestimate the power of the open-source - a user discovered a bug that caused worsened calibration of the fisheye setup.

Now, MC-Calib v1.2.1 fixes it and brings improved calibration of fisheye cameras

https://t.co/iw8lRa0SZf

[Alex Dailo @ukrdailo]

MC-Calib: a generic and robust calibration toolbox for multi-camera systems

TLDR: open source to calibrate multi-camera (perspective, fisheye, hybrid) systems with diverse configurations (overlapping, non-overlapping, converging).

GitHub: https://t.co/Gz5uuLXhog

[Alex Dailo @ukrdailo]

We have an #internship position in our ⁦@Qualcomm⁩ #XR Labs team in Amsterdam. Are you a (Ph.D.) student with experience and interest in research on efficient models for #3D (scene) reconstruction/understanding? https://t.co/K7bEIIeBUd

[Alex Dailo @ukrdailo]

We have an #internship position in our ⁦@Qualcomm⁩ #XR Labs team in Amsterdam. Are you a (Ph.D.) student with experience and interest in research on efficient models for #3D (scene) reconstruction/understanding? https://t.co/K7bEIIeBUd

[Maksym Andriushchenko @ ICML 🇰🇷 @maksym_andr]

🚨 I'm looking for a postdoc position to start in Fall 2024!

My most recent research interests are related to understanding foundation models (especially LLMs!), making them more reliable, and developing principled methods for deep learning.

More info: https://t.co/0LaOabpcxr

[Qualcomm Research & Technologies @QCOMResearch]

#ICCV2023 started today in Paris. These are @Qualcomm's #AI and #XR papers at the conference. Join us at the poster sessions to find out more:

[Alex Dailo @ukrdailo]

@mtmthh @ICCVConference Not this time @mtmthh

[Alex Dailo @ukrdailo]

Two papers from our team at @ICCVConference

1. DG-Recon: Depth-Guided Neural 3D Scene Reconstruction

TLDR: volumetric neural scene reconstruction with depth priors and multi-view feature fusion

[Qualcomm Research & Technologies @QCOMResearch]

4 days until #ICCV2023! Find out all about our #XR demos and come experience them at the @Qualcomm booth next week: "Depth-Guided Neural 3D Scene Reconstruction" and "Improved Self-supervised Depth Estimation on Reflective Surfaces" https://t.co/Zf8vqmLdC2

[Alex Dailo @ukrdailo]

Visit the posters at #ICCV2023 for more information and @Qualcomm booth for exciting demos

[Alex Dailo @ukrdailo]

2. 3D Distillation: Improving Self-Supervised Monocular Depth Estimation on Reflective Surfaces

TLDR: do the 3D fusion of the predicted depths, project them back to the corresponding frames, and use them as pseudo-labels to train the model