Mahdi Saleh | Technical University of Munich (TUM)

About Me

Hello! I’m Mahdi (also Mehdi) Saleh, a research scientist specializing in 3D computer vision at TUM, supervised by Federico Tombari and Nassir Navab. Currently, I am a research intern at Meta Reality Labs in Seattle, Washington. My research areas include 3D Object Pose, Hierarchical Representation, and Graph Neural Networks. I hold a Master’s degree in Biomedical Computing from TUM and a Bachelor’s in Electrical Control Engineering from IUST. In my leisure time, I enjoy computer graphics and design, web development, and game design.

Research Interests

3D computer vision and Machine Learning
6D Object pose estimation
Hierarchical representations and Graph Neural Networks
Vision Language Models

News

[2024] Paper accepted at ICRA 2024.
[2023] Paper accepted at ICCV 2023.
[2023] Paper accepted at MICCAI 2023 and 1 paper for ShapeMI MICCAI Workshop.
[2023] Paper accepted at CVPR 2023.
[2023] Starting a PhD research internship at Meta Reality Labs, Redmond, WA.
[2022] Won 2 awards at BOP 6D Object Pose 2022 challenge. Details
[2022] 1 paper accepted at IROS 2022, Kyoto, Japan.
[2022] 2 papers accepted at CVPR 2022, New Orleans, USA.
[2021] 1 paper accepted at NeurIPS 2021.
[2021] Starting a PhD research internship at Facebook Reality Labs, CA.
[2021] Recognized as an Outstanding Reviewer at CVPR 2021.
[2020] Accepted Oral paper at 3DV 2020.

Publications

ICRA

Physics-Encoded Graph Neural Networks for Deformation Prediction under Contact

Saleh, Mahdi; Sommersperger, Michael; Navab, Nassir; Tombari, Federico

IEEE International Conference on Robotics and Automation, 2024

Code Project Page

IROS

CloudAttention: Efficient Multi-Scale Attention Scheme For 3D Point Cloud Learning

Saleh, Mahdi; Wang, Yige; Navab, Nassir; Busam, Benjamin; Tombari, Federico

International Conference on Intelligent Robots and Systems, 2022

PDF

CVPR

Bending Graphs: Hierarchical Shape Matching using Gated Optimal Transport

Saleh, Mahdi; Wu, Shun-Cheng; Cosmo, Luca; Navab, Nassir; Busam, Benjamin; Tombari, Federico

IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022

PDF Code

CVPR

ZebraPose: Coarse to Fine Surface Encoding for 6DoF Object Pose Estimation

Su, Yongzhi; Saleh, Mahdi; Fetzer, Torben; Rambach, Jason; Navab, Nassir; Busam, Benjamin; Stricker, Didier; Tombari, Federico

IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022

PDF Code Winner of the BOP 2022 Challenge

NeurIPS

CoFiNet: Reliable Coarse-to-fine Correspondences for Robust PointCloud Registration

Yu, Hao; Li, Fu; Saleh, Mahdi; Busam, Benjamin; Ilic, Slobodan

Advances in Neural Information Processing Systems 34, 2021

PDF Code

3DV

Graphite: GRAPH-Induced feaTure Extraction for Point Cloud Registration

Saleh, Mahdi; Dehghani, Shervin; Navab, Nassir; Busam, Benjamin; Tombari, Federico

International Conference on 3D Vision (3DV). IEEE, 2020

PDF Oral Presentation

Student Projects

Supervising various Master’s theses and projects in the field of 3D computer vision, including topics like 3D shape completion, multimodal 3D understanding, and deformable shape matching.