Computer vision · RWTH Aachen

Kadir Yilmaz

I’m a PhD candidate in the Computer Vision Group at RWTH Aachen University.

My research aims to enable machines to understand the 3D world. I focus on simple, scalable architectures and rich representations for segmenting, tracking, and interpreting 3D scenes.

Kadir Yilmaz
Aachen, Germany

Research

All publications on Scholar
Volt partitions a 3D scene into patches and processes them with a vanilla Transformer for semantic segmentation.
ECCV 2026Long oral

Volume Transformer: Revisiting Vanilla Transformers for 3D Scene Understanding

Kadir Yilmaz*, Adrian Kruse*, Tristan Höfer, Daan de Geus, Bastian Leibe

A vanilla Transformer for 3D scenes, with volumetric patch tokens, global attention, and a training recipe that scales across datasets.

Paper ↗ Code ↗ Project ↗
BibTeX
@inproceedings{yilmaz2026Volt,
  title = {Volume Transformer: Revisiting Vanilla Transformers for 3D Scene Understanding},
  author = {Yilmaz, Kadir and Kruse, Adrian and Höfer, Tristan and de Geus, Daan and Leibe, Bastian},
  booktitle = {European Conference on Computer Vision (ECCV)},
  year = {2026}
}
DINO in the Room transfers visual foundation model features to 3D point cloud segmentation.
3DV 2026

DINO in the Room: Leveraging 2D Foundation Models for 3D Segmentation

Karim Knaebel*, Kadir Yilmaz*, Daan de Geus, Alexander Hermans, David Adrian, Timm Linder, Bastian Leibe

Bringing the knowledge of 2D foundation models into 3D segmentation through feature injection and distillation.

Paper ↗ Code ↗ Project ↗
BibTeX
@inproceedings{knaebel2026ditr,
  title = {DINO in the Room: Leveraging 2D Foundation Models for 3D Segmentation},
  author = {Knaebel, Karim and Yilmaz, Kadir and de Geus, Daan and Hermans, Alexander and Adrian, David and Linder, Timm and Leibe, Bastian},
  booktitle = {International Conference on 3D Vision (3DV)},
  year = {2026}
}
Interactive4D segments and tracks objects across a sequence of LiDAR scans using interactive clicks.
ICRA 2025

Interactive4D: Interactive 4D LiDAR Segmentation

Ilya Fradlin, Idil Esen Zulfikar, Kadir Yilmaz, Theodora Kontogianni, Bastian Leibe

Segmenting multiple objects across LiDAR scans in one interaction, with consistent object identities through time.

Paper ↗ Code ↗ Project ↗
BibTeX
@inproceedings{fradlin2025interactive4d,
  title = {Interactive4D: Interactive 4D LiDAR Segmentation},
  author = {Fradlin, Ilya and Zulfikar, Idil Esen and Yilmaz, Kadir and Kontogianni, Theodora and Leibe, Bastian},
  booktitle = {IEEE International Conference on Robotics and Automation (ICRA)},
  year = {2025}
}
Mask4Former assigns semantic labels and persistent instance identities to objects in a 4D LiDAR scene.
ICRA 2024

Mask4Former: Mask Transformer for 4D Panoptic Segmentation

Kadir Yilmaz, Jonas Schult, Alexey Nekrasov, Bastian Leibe

Joint segmentation and tracking of LiDAR point clouds using spatiotemporal instance queries.

Paper ↗ Code ↗ Project ↗
BibTeX
@inproceedings{yilmaz24mask4former,
  title = {Mask4Former: Mask Transformer for 4D Panoptic Segmentation},
  author = {Yilmaz, Kadir and Schult, Jonas and Nekrasov, Alexey and Leibe, Bastian},
  booktitle = {IEEE International Conference on Robotics and Automation (ICRA)},
  year = {2024}
}

* Equal contribution.

Further publications

Let’s connect

Get in touch.

For questions about my research or potential collaborations,
feel free to send me an email.

Computer Vision Group
RWTH Aachen University
Aachen, Germany