High-dimensional Deep Learning - ICPR Workshop 2020

High Dimensional Deeplearning

Held in conjunction with ICPR 2021; Milan, Italy, Januari 10th-15th 2021.

Higher dimensional data such as 3D, video, and simulation data are a leading edge of pattern recognition research. With the growth of prevalent application areas such as 3D games, self-driving automobiles, automobile and airplane design, health monitoring and sports activity training, a wide variety of new sensors and simulation techniques have allowed researchers to develop feature description models beyond 2D. In this workshop, we will present an overview and key insights into the state of the art of higher dimensional features from a wide variety of techniques including but not limited to deep learning and also traditional approaches. For example, numerous current pattern recognition methods are using 3D information from the sensor (e.g. KINECT, LIDAR, MRI, …) or are using 3D in modeling and understanding the 3D world. Although higher dimensional data enhance the performance of methods on numerous tasks, they can also introduce new challenges and problems. The higher dimensionality of the data often leads to more complicated structures which present additional problems in both extracting meaningful content and in adapting it for current machine learning algorithms.

Important dates

Submission deadline: October 10th 2020
Workshop author notification: November 10th 2020
Camera-ready submission: November 15th 2020
Finalized workshop program: December 1st 2020


  • 3D and 4D Deep Neural Architectures
  • 3D and 4D Deep Transfer Learning
  • 3D and 4D Deep Pattern Recognition
  • Learning 3D from single images
  • High-dimensional simulation data and modeling analytics
  • Fusion of high-dimensional features and classifiers
  • 3D and 4D Deep Features (e.g. temporal images/video, MRI, LIDAR, etc.)
  • Strengths and weaknesses of DL vs traditional approaches
  • Neuroevolutionary approaches
  • Optimization in latent space
  • Hyperparameter optimization
  • Geometric deep learning approaches
  • Very high dimensional deep learning (5D+)
  • Adapting low dimensionality methods to higher dimensions
  • Dimensionality reduction
  • Benchmarks with traditional approaches

Submission guidelines

All accepted papers are published in the Springer Workshop proceedings. There will be no printed version.
All papers should adhere to the Springer proceedings guidelines and templates.
Page limit:

  • Full paper (12-15 pages)
  • Short papers (6-8 pages)
  • No abstracts / demo and other submissions having fewer than 4 pages


Prof. Thomas Bäck

Head of the Natural Computing group
Leiden University

Prof. Michael Lew

Head of the Deep Learning and Computer Vision group
Leiden University

Dr. Markus Olhofer

Chief Scientist, HRI-EU

Dr. Bas van Stein

Deeplearning PostDoc
Leiden University

Program committee

  • Dr. Bart Thomee, senior researcher at Verify/Boston
  • Dr. Ahmed Nabil Belbachir, Chief Scientist, NORCE
  • Dr.-Ing. Stefan Menzel, Chief Scientist, Honda Research Institute Europe
  • Dr.-Ing. Steffen Limmer, Honda Research Institute Europe
  • Prof. Dr. Xiaohui Liu, Brunel University London
  • Dr. Yanming Guo, assistant professor at NUDT
  • Dr. Hao Wang, Sorbonne Université
  • Dr. Erwin Bakker, University of Leiden