The Imaging Systems Laboratory (ISL) at the Department of Electrical and Electronic Engineering, The University of Hong Kong (HKU), is delighted to announce that a research team led by Prof. Edmund Y. Lam has been awarded the Best Paper Award (Third Place) at the prestigious IEEE International Workshop on Machine Learning for Signal Processing (MLSP) 2024, held in London, UK. The award was granted for their paper titled “Ev-GS: Event-based Gaussian Splatting for Efficient and Accurate Radiance Field Rendering.”
The winning research, developed by ISL, marks the first computational neuromorphic imaging (CNI)-informed approach to infer 3D Gaussian splatting from monocular event cameras. This innovative method, named Ev-GS, addresses key challenges in radiance field rendering, such as detecting fast-moving objects and rendering scenes in low-light conditions. Unlike traditional frame-based methods, Ev-GS uses pure event-based signal supervision to achieve high-quality, blur-free, and efficient view synthesis. Experimental results reveal that Ev-GS significantly improves rendering speed and visual fidelity while reducing computational resource requirements.
This award underscores ISL’s leadership in computational imaging, radiance field rendering, and neuromorphic imaging, further enhancing HKU’s reputation in cutting-edge research in machine learning and signal processing.
For more details, please visit the project’s ArXiv page: https://arxiv.org/abs/2407.11343.