In April 2015, I was appointed as professor of “Computer Science in Engineering with a focus on engineering application systems” at HTW Berlin – University of Applied Sciences.
After studying mechanical engineering at the TU Ilmenau and automated production at ISMCM Paris, I held responsible positions in IT companies (Dassault Systèmes SOLIDWORKS Corp., eemagine Medical Imaging Solutions GmbH, PACE Aerospace Engineering and IT GmbH) for many years.

My research topics include 3D computer vision (depth sensors / 3D reconstruction / object recognition / similarity search) as well as deep learning in the field of 3D, Software design / architecture / engineering, and Mobile applications.


As part of HTW Berlin’s research strategy, I am involved in the Industrial Production research cluster

Research Areas

My research is directed towards the usage of 3D data originating from various depth sensing devices such as stereo cameras, time-of-flight cameras and LiDAR. My “3D Object Reconstruction & Recognition Lab” develops and evaluates methods for the fusion of depth maps to 3D models and for the instance level recognition of objects contained in large catalogs. In this context, we establish synthetic datasets for the training of deep-learning based methods as well as real-world validation data.

Depth Sensing Devices (Stereo Cameras, ToF, and LiDAR)

3D Object Reconstruction

Object Recognition based on Point Clouds

Deep Learning for 3D Data and Generation of Synthetic Datasets

Current Research Project

The Heatpump-Check offers a web portal for AI-powered review of quotes and for the optimal sizing of heat pumps to avoid oversizing. With financial support from the German Federal Environmental Foundation, the project launched on March 1 to prevent the oversizing of heat pumps in single-family homes and to achieve energy savings and cost reductions. We aim to make an initial version of the heat pump check available in early 2027.

Previous Research Projects

ServiceTechNavigator supports service technicians in navigating to the work site, finding their way around the plant premises and the permitted walking areas.

ELV.xD introduces AI-supported multidimensional ensemble learning methods for precise ad hoc spare parts recognition in an industrial environment.

With the KI-Lehr-Werkstatt Interdisziplinär (KIWI), a place for joint teaching, learning and application of AI technologies to issues from business and society is to be created at the HTW Berlin – University of Applied Sciences.

SparePartAssist helps service technicians to quickly find the right spare part on site.

Selected Publications

https://www.researchgate.net/profile/Frank-Neumann-6/research

News