High-fidelity Computational Photography Without AI

In at ime where deep learning techniques have become ubiquitous in computer vision and are increasingly common in image signal processing, real-time systems like live broadcasting remind us that classical image processing is far from irrelevant. In fact, when milliseconds matter, traditional algorithms remain highly relevant.

• At ImagingNext 2025, Jan Matthias Lüttgens, will explore what happens when you push image quality to the limit – without using AI. This talk is aimed to give  practical guidance on how to translate between image-quality and system specifications, at the example of a computational broadcast camera.

From Lab to Live Broadcast: How to achieve your image quality goals

Jan will give a walkthrough of the camera design process – starting with image quality as a design concept, then diving into practical examples from hardware measurement to quality benchmark and image tuning. Along the way, he will give some insights into broadcast, concluding with an assessment of deep-learning approaches in real-time applications.

About the Speaker: Jan Matthias Lüttgens

Jan Lüttgens received his Ph.D. in Physical Chemistry from Heidelberg University, Germany, where his research focused on strong light–matter coupling.

As a consultant at Eclipse Optics in Sweden, he worked on imaging and illumination optics, camera metrology, and full-stack camera simulation.
Currently, he is the Principal Image Engineer at Muybridge in Norway, leading the design of the imaging system architecture.

At ImagingNext, Jan combines deep technical expertise with real-world broadcast experience – making his session essential for anyone passionate about the future of high-quality, real-time imaging.

Why You Shouldn’t Miss This Session

Whether you’re working with imaging systems, video pipelines, or even AI, this session offers key takeaways

• what actually defines image quality
• how image quality benchmarks can guide camera system design
• how traditional methods are still relevant in image signal processing pipelines
• and how to approach tuning vs. training when building imaging systems.

You’ll also get a glimpse into the unique technical demands of broadcast production—and hear a few behind-the-scenes stories from the field.

If you’re interested in building faster, sharper, and more reliable imaging systems, this talk is for you.

Don’t miss your chance to see how vision applications are built efficiently — register now: https://framos.com/events/imaging-next-2025/