Jong-Beom Jeong

I'm a Post-Doc at Electronics and Telecommunications Research Institute (ETRI) in Daejeon, South Korea. I work in the media coding research section, and our team focuses on compression and representation of Gaussian splatting.

I did my PhD at Sungkyunkwan University (SKKU), where I was advised by Prof. Eun-Seok Ryu. I participated in the standardization of ISO/IEC 23090-12 MPEG immersive video (MIV), and conducted research including 360-degree video streaming, tile-based viewport-adaptive streaming, neural radiance fields (NeRF), and 3D Gaussian splatting (3DGS).

My research focuses on photorealistic volumetric media representation. Specifically, I'm interested in media compression and streaming to achieve low-latency and high-quality media.

Email  /  CV  /  LinkedIn  /  Scholar  /  ResearchGate  /  Gitlab  /  Github

News

• 01/2025 Our paper on GS compression is presented in Arxiv.
• 07/2024 Our paper on NeRF compression is accepted to MMSP 2024 (Oral).
• 02/2024 Our paper on MIV system is accepted to TOMM.
• 09/2023 Our paper on 360 VR streaming is accepted to VCIP 2023 (Oral).
• 05/2023 Our paper on INR compression is accepted to SPL.

Selected Publications

We use volumetric rendering with a sparse 3D feature grid and 2D feature planes to do real-time view synthesis.

Accounting for misalignment due to scene motion or calibration errors improves NeRF reconstruction quality.

We optimize a NeRF from scratch using a pretrained text-to-image diffusion model to do text-to-3D generative modeling.

Training a diffusion model on grid-based NeRFs lets you (conditionally) sample NeRFs.

NeRF lets us synthesize novel orthographic views that work well with pixel-wise algorithms for robotic manipulation.

A joint optimization framework for estimating shape, BRDF, camera pose, and illumination from in-the-wild image collections.

Miscellanea

Website template from Jon Barron.