PanoSAMic
Panoramic image segmentation using SAM
Panoramic image segmentation leveraging SAM feature encoding with dual-view fusion. Handles the distortion challenges of panoramic imagery by combining perspective and equirectangular representations.
Mahdi Chamseddine
Senior ML / Computer Vision Engineer
ML researcher and engineer with 7+ years building deep learning systems for computer vision and 3D scene perception: foundation models, semantic segmentation, multi-sensor fusion, and end-to-end pipelines from distributed training to deployment. PhD in Computer Science at RPTU Kaiserslautern-Landau, expected 2026. Background in electrical engineering, automation, and control systems.
Projects
SAM-based Video Annotation Platform
Multi-user video annotation platform built on SAM
Developed, shipped, and maintain a SAM-based video annotation platform — multi-user authentication, user management, and a Gradio client-facing UI — delivered to an industry partner for production use.
Internal tool — code not public
PyStruct3D & OpenBIMxD
Open-source Python libraries for scan-to-BIM workflows
Two complementary open-source libraries for automated scan-to-BIM pipelines — point cloud processing and geometry fitting, through to structured IFC file generation.
CaRaCTO-3D
Camera-radar calibration to 3D scene reconstruction
Robust camera-radar extrinsic calibration method using triple constraint optimization, extended to full 3D scene reconstruction from fused sensor data. Best Industrial Paper Award at ICPRAM 2024.
ToF-360
Panoramic time-of-flight RGB-D dataset
Contributed to creating a panoramic ToF RGB-D dataset for single-capture indoor semantic 3D reconstruction. Published at CVPR 2025 Workshops.
Publications
M. Chamseddine, D. Stricker, and J. Rambach, “PanoSAMic: Panoramic Image Segmentation from SAM Feature Encoding and Dual View Fusion,” ICPR, 2026 (to appear).
M. Chamseddine, F. Kaufmann, M. Schellen, C. Glock, D. Stricker, and J. Rambach, “BIMStruct3D: A Fully Automated Hybrid Learning Scan-to-BIM Pipeline with Integrated Topology Refinement,” EC3, 2026 (to appear).
S. Inuganti, H. Kanayama, K. Shimizu, M. Chamseddine, S. Yokota, D. Stricker, and J. Rambach, “JOPP-3D: Joint Open Vocabulary Semantic Segmentation on Point Clouds and Panoramas,” arXiv preprint, 2026.
M. Chamseddine, J. Rambach, and D. Stricker, “CaRaCTO-3D: From Camera-Radar Calibration to Scene Reconstruction,” SN Computer Science, 2025.
H. Kanayama, M. Chamseddine, S. Guttikonda, S. Okumura, S. Yokota, D. Stricker, and J. Rambach, “ToF-360 — A Panoramic Time-of-Flight RGB-D Dataset for Single Capture Indoor Semantic 3D Reconstruction,” CVPR Workshops, 2025.
M. Chamseddine, J. Rambach, and D. Stricker, “CaRaCTO: Robust Camera-Radar Extrinsic Calibration with Triple Constraint Optimization,” ICPRAM, 2024. Best Industrial Paper
F. Kaufmann, M. Chamseddine, S. Guttikonda, C. Glock, D. Stricker, and J. Rambach, “Ontology-Based Semantic Labeling for RGB-D and Point Cloud Datasets,” EC3, 2023.
I. Brishtel, S. Krauss, M. Chamseddine, J. R. Rambach, and D. Stricker, “Driving Activity Recognition Using UWB Radar and Deep Neural Networks,” Sensors, 2023.
M. Chamseddine, J. Rambach, D. Stricker, and O. Wasenmüller, “Ghost Target Detection in 3D Radar Data Using Point Cloud Based Deep Neural Network,” ICPR, 2021.