VLG | Computer Vision and Learning Group

Basic Information

I am a postdoctoral fellow at the AI Center of ETH Zurich working with Prof. Dr. Siyu Tang and Prof. Dr. Otmar Hiliges . Previously, I finished my Ph.D. in Computer Vision, Machine Learning and 3D Scene Understanding in the Computer Vision Group of Prof. Dr. Bastian Leibe at RWTH Aachen University.

We have several exciting projects in the field of 3D scene understanding, human-scene interaction and synthesis available. To apply for a bachelor/master/semester thesis, send me an informal email.

Social

Publications

Human3D: 3D Segmentation of Humans in Point Clouds with Synthetic Data

Conference: International Conference on Computer Vision (ICCV 2023)

Authors:Ayça Takmaz^, Jonas Schult^, Irem Kaftan, Mertcan Akçay, Bastian Leibe, Robert Sumner, Francis Engelmann, Siyu Tang
We propose the first multi-human body-part segmentation model, called Human3D 🧑‍🤝‍🧑, that directly operates on 3D scenes. In an extensive analysis, we validate the benefits of training on synthetic data on multiple baselines and tasks.

Project PDF BibTeX

Box2Mask: Weakly Supervised 3D Semantic Instance Segmentation Using Bounding Boxes

Conference: International Conference on 3D Vision (3DV 2021) oral presentation

Authors:Julian Chibane, Francis Engelmann, Tuan Anh Tran, Gerard Pons-Moll
Are 3D bounding box annotations suitable to train dense 3D semantic instance segmentation? Yes!

Project PDF Code BibTeX

Mix3D: Out-of-Context Data Augmentation for 3D Scenes

Conference: International Conference on 3D Vision (3DV 2021) oral presentation

Authors:Alexey Nekrasov, Jonas Schult, Or Litany, Bastian Leibe, Francis Engelmann
Mix3D is a data augmentation technique for segmenting large-scale 3D scenes.

Project PDF Code BibTeX

Points2Objects: From Points to Multi-Object 3D Reconstruction

Conference: Computer Vision and Pattern Recognition (CVPR), 2021

Authors:Francis Engelmann, Konstantinos Rematas, Bastian Leibe, Vittorio Ferrari
We propose a method to detect and reconstruct multiple 3D objects from a single RGB image.

Project PDF Code BibTeX

3D-MPA: Multi Proposal Aggregation for 3D Semantic Instance Segmentation

Conference: Computer Vision and Pattern Recognition (CVPR), 2020

Authors:Francis Engelmann, Martin Bokeloh, Alireza Fathi, Bastian Leibe, Matthias Nießner
We present 3D-MPA, a method for instance segmentation on 3D point clouds.

Project PDF Code BibTeX

DualConvMesh-Net: Joint Geodesic and Euclidean Convolutions on 3D Meshes

Conference: Computer Vision and Pattern Recognition (CVPR), 2020 oral presentation

Authors:Jonas Schult, Francis Engelmann, Theodora Kontogianni, Bastian Leibe
We propose DualConvMesh-Nets (DCM-Net) a family of deep hierarchical convolutional networks over 3D geometric data that combines two types of convolutions.

Project PDF Code BibTeX

Basic Information

Social

Publications

Human3D: 3D Segmentation of Humans in Point Clouds with Synthetic Data

Box2Mask: Weakly Supervised 3D Semantic Instance Segmentation Using Bounding Boxes

Authors:Julian Chibane, Francis Engelmann, Tuan Anh Tran, Gerard Pons-Moll Are 3D bounding box annotations suitable to train dense 3D semantic instance segmentation? Yes!

Mix3D: Out-of-Context Data Augmentation for 3D Scenes

Authors:Alexey Nekrasov, Jonas Schult, Or Litany, Bastian Leibe, Francis Engelmann Mix3D is a data augmentation technique for segmenting large-scale 3D scenes.

Points2Objects: From Points to Multi-Object 3D Reconstruction

Authors:Francis Engelmann, Konstantinos Rematas, Bastian Leibe, Vittorio Ferrari We propose a method to detect and reconstruct multiple 3D objects from a single RGB image.

3D-MPA: Multi Proposal Aggregation for 3D Semantic Instance Segmentation

Authors:Francis Engelmann, Martin Bokeloh, Alireza Fathi, Bastian Leibe, Matthias Nießner We present 3D-MPA, a method for instance segmentation on 3D point clouds.

DualConvMesh-Net: Joint Geodesic and Euclidean Convolutions on 3D Meshes

Authors:Jonas Schult*, Francis Engelmann*, Theodora Kontogianni, Bastian Leibe We propose DualConvMesh-Nets (DCM-Net) a family of deep hierarchical convolutional networks over 3D geometric data that combines two types of convolutions.

Authors:Julian Chibane, Francis Engelmann, Tuan Anh Tran, Gerard Pons-Moll
Are 3D bounding box annotations suitable to train dense 3D semantic instance segmentation? Yes!

Authors:Alexey Nekrasov, Jonas Schult, Or Litany, Bastian Leibe, Francis Engelmann
Mix3D is a data augmentation technique for segmenting large-scale 3D scenes.

Authors:Francis Engelmann, Konstantinos Rematas, Bastian Leibe, Vittorio Ferrari
We propose a method to detect and reconstruct multiple 3D objects from a single RGB image.

Authors:Francis Engelmann, Martin Bokeloh, Alireza Fathi, Bastian Leibe, Matthias Nießner
We present 3D-MPA, a method for instance segmentation on 3D point clouds.

Authors:Jonas Schult, Francis Engelmann, Theodora Kontogianni, Bastian Leibe
We propose DualConvMesh-Nets (DCM-Net) a family of deep hierarchical convolutional networks over 3D geometric data that combines two types of convolutions.