Karim Knaebel

I’m a PhD student in Computer Vision at RWTH Aachen University, advised by Prof. Bastian Leibe. My current focus is monocular 3D reconstruction; previously I worked on 3D scene understanding.

Previously Karim Abou Zeid. Outside research, I enjoy motorcycles, nature, and building software tools.

Papers

SurGe: Improved Surface Geometry in Point Maps

Karim Knaebel, Gonzalo Martin Garcia, Christian Schmidt, Ilya Fradlin, Lucas Nunes, Daan de Geus, Bastian Leibe

arXiv

arXiv / pdf / project / code / demo / bibtex

Improved local surface geometry in point maps via a Neighborhood Attention Decoder (NAD).
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

3DV 2026

arXiv / pdf / project / code / bibtex

Injecting and distilling DINOv2 features into 3D segmentation models for stronger indoor and outdoor performance.
Fine-Tuning Image-Conditional Diffusion Models is Easier than You Think

Gonzalo Martin Garcia, Karim Knaebel, Christian Schmidt, Daan de Geus, Alexander Hermans, Bastian Leibe

WACV 2025

arXiv / pdf / project / code / bibtex

Single-step fine-tuning of diffusion models for depth and normal estimation with strong generalization.
Benchmarks and Challenges in Pose Estimation for Egocentric Hand Interactions with Objects

Zicong Fan, Takehiko Ohkawa, Linlin Yang, Nie Lin, Zhishan Zhou, Shihao Zhou, Jiajun Liang, Zhong Gao, Xuanyang Zhang, Xue Zhang, Fei Li, Zheng Liu, Feng Lu, Karim Knaebel, Bastian Leibe, Jeongwan On, Seungryul Baek, Aditya Prakash, Saurabh Gupta, Kun He, Yoichi Sato, Otmar Hilliges, Hyung Jin Chang, Angela Yao

ECCV 2024

arXiv / pdf / bibtex

Analysis of HANDS23 challenge results for egocentric hand-object pose estimation, highlighting key challenges and effective modeling strategies.
JointTransformer

Karim Knaebel, Bastian Leibe

ICCV 2023 HANDS Workshop

code / bibtex

Winner of the HANDS ARCTIC Challenge with joint reasoning over hands and objects.
Point2Vec for Self-Supervised Representation Learning on Point Clouds

Karim Knaebel, Jonas Schult, Alexander Hermans, Bastian Leibe

GCPR 2023

arXiv / pdf / project / code / bibtex

Self-supervised pretraining on point clouds with strong downstream performance on classification and segmentation.

Code

turm — TUI for the Slurm Workload Manager
cfgx — Python-first config loader with parent chaining and CLI overrides
specbuild — Registry-based object builder for nested configuration dictionaries
trainloop — Minimal PyTorch training loop with hooks for logging and checkpoints
itar — Tar file index for constant-time member access
arxivit — LaTeX source cleaner with optional image resizing for arXiv uploads

Legacy work:

Iron — Modern weightlifting workout tracker for iOS, written in SwiftUI
Phonograph — Material Design music player for Android

Contact

Email: [email protected]