Image: Ralf A. Niggemann.
Image: Ralf A. Niggemann.

Wetzlar Network

Computational Imaging

Wetzlar Network organized an international expert panel on the subject of computational imaging which took place in Wetzlar on 22nd and 23rd October, 2019. Owing to the extremely positive feedback, the panel might be repeated next year.

It is a well-known fact that Photoshop allows to manipulate, optimize or even butcher digital photographs. But if the methods of image processing are based almost completely on algorithms and statistical approaches to a problem like “Deep Learning” and artificial intelligence, they might rock the foundations of many a cherished habit. By now, these methods circulating under the label “Computational Imaging” have become so sophisticated that they exceed previous quality standards by far. They could change the future of photography and microscopy as well as the field of three-dimensional reconstruction.

One thing is certain: Computational Imaging is a hot topic and of far-reaching relevance which became more than evident in the course of a two-day expert panel on 22nd and 23rd October, 2019. International experts and more than 50 intrigued participants attended the event in Leitz Park Wetzlar to discuss the prospects and opportunities of Computational Imaging.

City councilman Norbert Kortlüke bid the guests a warm welcome in the City of Optics. Speakers and participants were also greeted by Ralf Niggemann of Wetzlar Network and Prof. Dr. Sangam Chatterjee of the Justus Liebig University, Gießen. And Dr. Andreas Kaufmann who opened the panel left no doubt that he was closely following the latest developments in this field not only as the host of the event but above all as the chairman of the supervisory board of Leica Camera AG.

Dr. Shwetadwip Chowdhury of the University of California Berkeley and Dr. E. Y. Peng of Stanford University talked about their research work in the USA. Dr. Valentin Volchkov of Max Planck Institute for Intelligent Systems in Tübingen presented a method which allows to calculate the qualities of a lens (so-called MTF) based on an image. Presentations were held on minimizing the image noise without impairing the image quality (Prof. Dr. Stefan Roth of Darmstadt University of Technology), and a new camera family was introduced which already has neuronal networks “on board” (Dr. Robert-Alexander Windberger of IDS Imaging Development Systems GmbH).

Prof. Dr. Paolo Favaro of the University Bern presented an algorithm which can be used to eliminate the blur of moving objects. And Prof. Yoann Altmann of the Heriot-Watt University Edinburgh talked, among other things, about a method which allows to look “around the corner” (non-line-of-sight – nlos) with the aid of the appropriate algorithms.

The top-class speakers demonstrated formidably what can already be done in the field of image generation and what will be possible in future. “Every talk and every discussion during the last two days has been an asset,” Benjamin Dück of Leica Camera says summing it up - and he expresses his hope “that word about the quality of Computational Imaging will spread.” The movers and shakers of Wetzlar Network, Ralf Niggemann and André Noack, are willing to do a lot to achieve this end and intend to get even more companies from the region involved. “The feedback was extremely positive,” André Noack says delightedly, “and there are many indications that a second Computational Imaging Expert Panel will take place next year.” Particularly since the subject is certainly not a nine-day wonder but could keep the optics region occupied in the long term. The further developments will be interesting to see.