Mohamed Sayed


Hello! I'm Mohamed, a Senior Research Scientist at Niantic, Inc.

I am excited about 3D reconstruction and real-world robust interactive vision systems. My recent projects include depth estimation and reconstruction (SimpleRecon, DoubleTake), soft occlusions for AR effects, and primitive geometry estimation for scenes. I also designed and built the majority of LookOut, an interactive filming-copilot system for automatically pointing a camera during filming.

Publications

DoubleTake: Geometry Guided Depth Estimation

Mohamed Sayed, Filippo Aleotti, Jamie Watson, Zawar Qureshi, Guillermo Garcia-Hernando, Gabriel Brostow, Sara Vicente, Michael Firman

European Conference on Computer Vision - ECCV 2024

Project Page, Paper, Code, Video, Bibtex

Rapid Sketch-Based 3D City Modeling

Gizem Esra Unlu, Mohamed Sayed, Yulia Gryaditskaya, Gabriel Brostow

European Conference on Computer Vision - ECCV 2024

Bibtex

AirPlanes Accurate Plane Estimation via 3D-Consistent Embeddings

Jamie Watson, Filippo Aleotti, Mohamed Sayed, Zawar Qureshi, Oisin Mac Aodha, Gabriel Brostow, Michael Firman, Sara Vicente

Computer Vision and Pattern Recognition - CVPR 2024

Project Page, Paper, Code, Video, Bibtex

An Empirical Study of the Generalization Ability of Lidar 3D Object Detectors to Unseen Domains

George Eskandar, Chongzhe Zhang, Abhishek Kaushik, Karim Guirguis, Mohamed Sayed, Bin Yang

Computer Vision and Pattern Recognition - CVPR 2024

Paper

Don’t Look Now: Audio/Haptic Guidance for 3D Scanning of Landmarks

Jessica Van Brummelen, Liv Piper Urwin, Oliver James Johnston, Mohamed Sayed, Gabriel Brostow

CHI Conference on Human Factors in Computing Systems - CHI 2024

Paper, Video

Virtual Occlusions Through Implicit Depth

Jamie Watson, Mohamed Sayed, Zawar Qureshi, Gabriel Brostow, Sara Vicente, Oisin Mac Aodha, Michael Firman

Computer Vision and Pattern Recognition - CVPR 2023

Project Page, Paper, Code, Video, Bibtex

SimpleRecon: 3D Reconstruction Without 3D Convolutions

Mohamed Sayed, John Gibson, Jamie Watson, Victor Prisacariu, Michael Firman, Clément Godard

European Conference on Computer Vision - ECCV 2022

Project Page, Paper, Code, Video, Bibtex, Media: Niantic Labs Blog, Marktechpost Article

LookOut! Interactive Camera Gimbal Controller for Filming Long Takes

Mohamed Sayed, Robert Cinca, Enrico Costanza, Gabriel Brostow

Transactions on Graphics - ToG 2022, SIGGRAPH 2022

Project Page, Paper, 30s Fast Forward, Video, Filmed Scenes, Bibtex

Interactive Sketching of Mannequin Poses

Gizem Esra Ünlü, Mohamed Sayed, Gabriel Brostow

International Conference on 3D Vision - 3DV 2022

Project Page, Paper, Video, Bibtex

Improved Handling of Motion Blur in Online Object Detection

Mohamed Sayed, Gabriel Brostow

Computer Vision and Pattern Recognition - CVPR 2021

Project Page, Paper, Video, Code, Bibtex
Work Experience

June 2024 -
London

Senior Research Scientist, Niantic, Inc.

April 2023 - June 2024
London

Research Scientist, Niantic, Inc.

Aug 2022 - Feb 2023,
London

Part-time Researcher, Niantic, Inc.

Production work for SimpleRecon

April - July 2022,
Zurich

Research Intern, Disney Research | Studios

Controllable Super Resolution and Diffusion Models

May - Dec 2021,
London

Research Intern, Niantic, Inc.

Depth Estimation and 3D Reconstruction Project - Work accepted at ECCV 2022 and patent pending.

Spring 2017,
Cairo

Software Intern, Valeo

Developed automation software for hardware integration testing. Reduced regression testing time significantly.

Summer 2016,
Cairo

IT Intern, Mercedes-Benz Egypt

Recognized by the CEO for reporting and fixing errors and loopholes in the employee time tracking system.

Education

September 2018 - March 2023

PhD in Computer Science, Computer Vision (University College London)

Supervised by Prof. Gabriel Brostow.

September 2017 - September 2018

MSc Computer Graphics, Vision, and Imaging (University College London)

Distinction and Dean's List. Thesis: "Scripted Camera Control Through Visual Tracking." Thesis mark: 91/100.

September 2012 - May 2017

Bachelor of Electrical Engineering, Communications & Computer Engineering (Cairo University)

Distinction with honors. Thesis: Improving the performance of a widely used Mentor tool with Process Mining.

Academic Peer Reviewing

Reviewer

Conferences: ECCV ('24), CVPR ('23, '24), ICCV ('23), SIGGRAPH ('23), SIGGRAPH Asia ('23), BMVC ('22). Journals: TPAMI ('22), IJCV ('22)

Teaching

University College London

Voted Best Teaching Assistant (TA) in Computer Science (2018/2019)

Machine Vision TA - 2018, 2019, 2020 (head)

Computer Graphics TA - 2020

Image Processing TA - 2018, 2019

Computational Photography and Capture TA - 2019, 2020

Other Projects

Melanoma Classification (2016)

Built a classifier for diagnosing melanoma skin lesions through dermoscopic images.

Supervisor: Prof. Tawfik Ismail.

Telescope Guidance System (2016)

Amateur astrophotography with an 8-inch Newtonian telescope. Prototyped a telescope guidance system (hardware and software) with an Arduino and the ASCOM platform. Images at astrobin.com/users/DexPrime/.

“YallaCode” Android App (2015)

Led a team of four to develop the app to teach children how to build basic computer programs. The app featured a built in “block statement” compiler.

Musical Glove (2013)

Built a “Musical Glove” using Arduino that tracks hand movements and alters musical notes on a synthesizer through a virtual MIDI driver.

Skills

Languages

Python, MATLAB, C++, C#, Assembly

Deep Learning Libraries

Pytorch, Tensorflow

Useful Tools

COLMAP, Inkscape, Adobe Photoshop, Adobe After Effects, Google Cloud Compute

Hardware

Arduino, Soldering, 3D Printing, Real-time Control, Serial Communication

Hobbies

Photography, Cooking, Running, Model Making