Paul Henderson

I am a Lecturer (roughly equivalent to Assistant Professor) in Machine Learning at the University of Glasgow. My research focuses on building machines that understand the visual world with minimal supervision, learning aspects of its structure such as 3D geometry and decomposition into objects. This work lies at the intersection of machine learning, computer vision, and computer graphics. I also work on applications of machine learning and computer vision in the physical and life sciences.

Prospective PhD students: I am currently accepting PhD students to supervise. There are studentships available for strong candidates. Please first check that your research interests align with mine (read this page; read some of my papers). If so, send me an email with your CV, transcripts and a short research proposal (2–3 pages) mentioning why you want to work in my group in particular, and what skills make you suited. Note that I expect students to produce at least two 'good' papers during their PhD (CVPR, NeurIPS, etc.). Applicants should have a strong background in coding and maths (e.g. probability, linear algebra).

Research

RenderDiffusion: Image Diffusion for 3D Reconstruction, Inpainting and Generation

Titas Anciukevičius, Zexiang Xu, Matthew Fisher, Paul Henderson, Hakan Bilen, Niloy J. Mitra and Paul Guerrero

CVPR 2023 (new!)

Preprint – Bibtex

@article{anciukevicius2022renderdiffusion,
	title        = {{RenderDiffusion}: Image Diffusion for {3D} Reconstruction, Inpainting and Generation},
	author       = {Titas Anciukevicius and Zexiang Xu and Matthew Fisher and Paul Henderson and Hakan Bilen and Mitra, Niloy J. and Paul Guerrero},
	year         = 2023,
	journal      = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)}
}

Simulating analogue film damage to analyse and improve artefact restoration on high-resolution scans

Daniela Ivanova, John Williamson and Paul Henderson

Eurographics 2023 / Computer Graphics Forum (new!)

Paper – Bibtex

@article{ivanova23analogue,
	title        = {Simulating analogue film damage to analyse and improve artefact restoration on high-resolution scans},
	author       = {Daniela Ivanova and John Williamson and Paul Henderson},
	year         = 2023,
	journal      = {Computer Graphics Forum (Proc. Eurographics 2023)},
}

Unsupervised Causal Generative Understanding of Images

Titas Anciukevicius, Patrick Fox-Roberts, Edward Rosten and Paul Henderson

NeurIPS 2022

Full Paper – Short Paper (ICML SCIS Workshop) – Bibtex

@inproceedings{anciukevicius22neurips,
  title={Unsupervised Causal Generative Understanding of Images},
  author={Titas Anciukevi\v{c}ius and Patrick Fox-Roberts and Edward Rosten and Paul Henderson},
  booktitle={Advances in Neural Information Processing Systems (NeurIPS) 35},
  year={2022}
}

Determining band structure parameters of two-dimensional materials by deep learning

Paul Henderson, Areg Ghazaryan, Alexander A. Zibrov, Andrea F. Young and Maksym Serbyn

arXiv 2022

Preprint – Bibtex

@misc{henderson22corr,
  title={Determining band structure parameters of two-dimensional materials by deep learning},
  author={Paul Henderson and Areg Ghazaryan and Zibrov, Alexander A. and Young, Andrea F. and Maksym Serbyn},
  howpublished={arXiv:2210.06310},
  year={2022}
}

Learning to Predict Keypoints and Structure of Articulated Objects without Supervision

Titas Anciukevicius, Paul Henderson and Hakan Bilen

ICPR 2022

Paper – Bibtex

@inproceedings{anciukevicius22icpr,
  title={Learning to Predict Keypoints and Structure of Articulated Objects without Supervision},
  author={Titas Anciukevi\v{c}ius and Paul Henderson and Hakan Bilen},
  booktitle={International Conference on Pattern Recognition (ICPR)},
  year={2022}
}

Unsupervised Video Prediction from a Single Frame by Estimating 3D Dynamic Scene Structure

Paul Henderson, Christoph Lampert and Bernd Bickel

CoRR 2021

Project Page – Paper – Bibtex

@misc{henderson21corr,<br />
            title={Unsupervised Video Prediction from a Single Frame by Estimating {3D} Dynamic Scene Structure},<br />
            author={Henderson, Paul and Lampert, Christoph H. and Bickel, Bernd},<br />
            howpublished={arXiv:2106.09051},<br />
            year={2021}<br />
        }

Unsupervised object-centric video generation and decomposition in 3D

Paul Henderson and Christoph Lampert

NeurIPS 2020

Project Page – Paper – Code – Bibtex

@inproceedings{henderson20neurips,
            title={Unsupervised object-centric video generation and decomposition in {3D}},
            author={Henderson, Paul and Lampert, Christoph H.},
            booktitle={Advances in Neural Information Processing Systems (NeurIPS) 33},
            year={2020}
    }

Computational Design of Cold Bent Glass Façades

Konstantinos Gavriil, Ruslan Guseinov, Jesús Pérez, Davide Pellis, Paul Henderson, Florian Rist, Helmut Pottmann and Bernd Bickel

SIGGRAPH Asia 2020

Project Page – Paper – arXiv – Bibtex

@article{gavriil20sa,
  title={Computational Design of Cold Bent Glass Fa{\c c}ades},
  author={Gavriil, Konstantinos and Guseinov, Ruslan and P{\'e}rez, Jes{\'u}s and Pellis, Davide and Henderson, Paul and Rist, Florian and Pottmann, Helmut and Bickel, Bernd},
  journal={ACM Transactions on Graphics (SIGGRAPH Asia 2020)},
  year={2020},
  month = {12},
  volume = {39},
  number = {6}
  articleno = {208},
  numpages = {16}
}

Leveraging 2D Data to Learn Textured 3D Mesh Generation

Paul Henderson, Vagia Tsiminaki and Christoph Lampert

CVPR 2020 (oral)

Paper – Code – Bibtex

@inproceedings{henderson20cvpr,
  title={Leveraging {2D} Data to Learn Textured {3D} Mesh Generation},
  author={Paul Henderson and Vagia Tsiminaki and Christoph Lampert},
  booktitle={IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2020}
}

Object-Centric Image Generation with Factored Depths, Locations, and Appearances

Titas Anciukevicius, Christoph Lampert and Paul Henderson

CoRR 2020

Paper – ICML 2020 workshop presentation – Bibtex

@misc{anciukevicius20corr,
  title={Object-Centric Image Generation with Factored Depths, Locations, and Appearances},
  author={Titas Anciukevi\v{c}ius and Christoph Lampert and Paul Henderson},
  howpublished={arXiv:2004.00642},
  year={2020}
}

Learning Single-Image 3D Reconstruction by Generative Modelling of Shape, Pose and Shading

Paul Henderson and Vittorio Ferrari

IJCV 2019

Paper – arXiv – Code for differentiable renderer – Bibtex

@article{henderson19ijcv,
  title={Learning Single-Image {3D} Reconstruction by Generative Modelling of Shape, Pose and Shading},
  author={Henderson, Paul and Ferrari, Vittorio},
  journal={International Journal of Computer Vision (IJCV)},
  year={2019},
  doi={10.1007/s11263-019-01219-8}
}

Learning to Generate and Reconstruct 3D Meshes with only 2D Supervision

Paul Henderson and Vittorio Ferrari

BMVC 2018 (oral)

Paper – Code for differentiable renderer – Bibtex

@inproceedings{henderson18bmvc,
  title={Learning to Generate and Reconstruct {3D} Meshes with only {2D} Supervision},
  author={Henderson, Paul and Ferrari, Vittorio},
  booktitle={British Machine Vision Conference (BMVC)},
  year={2018}
}

Automatic Generation of Constrained Furniture Layouts

Paul Henderson, Kartic Subr and Vittorio Ferrari

CoRR 2017

Paper – Bibtex

@misc{henderson17corr,
  title={Automatic Generation of Constrained Furniture Layouts},
  author={Paul Henderson and Kartic Subr and Vittorio Ferrari},
  howpublished={arXiv:1711.10939},
  year={2017}
}

End-to-end Training of Object Class Detectors for Mean Average Precision

Paul Henderson and Vittorio Ferrari

ACCV 2016

Paper – Bibtex

@inproceedings{henderson16accv,
  title={End-to-End Training of Object Class Detectors for Mean Average Precision},
  author={Henderson, Paul and Ferrari, Vittorio},
  booktitle={Asian Conference on Computer Vision (ACCV)},
  year={2016},
  pages={198--213}
}

Automatically Selecting Inference Algorithms for Discrete Energy Minimisation

Paul Henderson and Vittorio Ferrari

ECCV 2016

Paper – arXiv – Bibtex

@inproceedings{henderson16eccv,
  title={Automatically Selecting Inference Algorithms for Discrete Energy Minimisation},
  author={Henderson, Paul and Ferrari, Vittorio},
  booktitle={European Conference on Computer Vision (ECCV)},
  year={2016},
  pages={235--252}
}

Profile

I completed a BA in Mathematics at the University of Cambridge in 2009, followed by an MSc in Informatics at the University of Edinburgh in 2010. I then worked at Blackford Analysis for four years, on research and development for medical imaging applications. I completed my PhD in 2018 at the University of Edinburgh in the CALVIN group supervised by Vittorio Ferrari (see here for my thesis). I spent three years as a postdoc in the Computer Vision and Machine Learning Group of Christoph Lampert at ISTA, the Institute of Science and Technology Austria. Since January 2022, I hold the post of Lecturer in Machine Learning, in the School of Computing Science at the University of Glasgow. My 'official' (not-so-interesting) homepage is here.

My curriculum vitae is available here.