Colloquiums and Conferences

From Camera to AI: Challenges in Visual Content Creation

Time: 2025-11-03 12:08:00

Title: From Camera to AI: Challenges in Visual Content Creation
Speaker: Prof. Guoping Qiu (Invited by Prof. Weibin Wang)
Affiliation: University of Nottingham Ningbo China / School of Computer Science, University of Nottingham
Time: November 3, 10:00–11:00
Venue: Lecture Hall, Room 111, West Building, School of Mathematical Sciences

Speaker Biography:
    Guoping Qiu is a Chair Professor in the School of Computer Science at the University of Nottingham and Vice-Provost of the University of Nottingham Ningbo China. He has long focused on neural networks and image processing. Early in his career he applied unsupervised competitive neural networks to image feature extraction, proposing one of the early approaches to representation learning; well before deep learning became popular, he proposed learning-based super-resolution and the use of neural networks to remove compression artefacts. In HDR imaging, he and his team introduced key techniques such as tone mapping that changed the way HDR content is processed and displayed, and the resulting technologies were commercialised in award-winning software including HDR Darkroom and Fotor, used by hundreds of millions of people worldwide. As Chief Scientist of Everimaging, the developer of HDR Darkroom and Fotor, he continues to apply imaging technology to real-world problems. His current research centres on deep learning, vision–language modelling and large vision models applied to key challenges in digital photography. He has published more than 300 research papers and holds a number of US and European patents, and is ranked in the top 0.5% of scientists in the artificial intelligence and image processing subfield of Stanford University's list of the world's top 2% scientists.

Abstract:
    Vision plays a fundamental role in human perception, learning and cognition, and more than 80% of our interaction with the world relies on it. It is therefore unsurprising that computational visual content creation, from recording the beauty of nature with a camera to generating entirely virtual scenes with artificial intelligence, has advanced remarkably in recent decades. Nevertheless, key challenges remain. Conventional digital photography still struggles with the mismatch between the enormous dynamic range of natural scenes and the limitations of display and printing media, which restricts visual fidelity. At the same time, making AI-generated content (AIGC) obey the physical laws of the real world while retaining creative flexibility remains an open problem. Moreover, even when theoretical breakthroughs are achieved, turning them into usable tools that meet real user needs is a demanding task. In this talk, we discuss the key technical difficulties in natural imaging and AI-driven visual creation, present recent solutions that bridge theory and practice, and demonstrate an AI-driven digital creation platform designed to help users produce high-quality visual content with ease.