Georg Hess

Industrial PhD Student @zenseact. Deep learning for multi-object tracking.

[MrNeRF @janusch_patas]

SplatAD: Real-Time Lidar and Camera Rendering with 3D Gaussian Splatting for Autonomous Driving

Contributions:
• We propose the first method for efficient lidar rendering using 3D Gaussians, introducing custom CUDA-accelerated algorithms for rasterizing sparse point clouds in spherical coordinates.

• We introduce the first 3DGS method capable of rendering both camera and lidar from a unified representation, enabling accelerated scaling of novel view synthesis for automotive applications.

• We present effective techniques for realistic sensor modeling using 3D Gaussians, enabling accurate handling of rolling shutter, lidar intensity, ray dropping, and variations in sensor appearance.

• Through extensive evaluation on three popular automotive datasets, we demonstrate state-of-the-art results across all benchmarks, validating our method’s effectiveness and generalizability.

[Manuel Dahnert @manuel_dahnert]

Super happy to present our #NeurIPS paper 𝐂𝐨𝐡𝐞𝐫𝐞𝐧𝐭 𝟑𝐃 𝐒𝐜𝐞𝐧𝐞 𝐃𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧 𝐅𝐫𝐨𝐦 𝐚 𝐒𝐢𝐧𝐠𝐥𝐞 𝐑𝐆𝐁 𝐈𝐦𝐚𝐠𝐞 in Vancouver.

Come to our poster #2804 on Wednesday 11am - 2pm in East Exhibit Hall A-C and say hi if you want to learn more about 3D Scene Diffusion!

See you tomorrow!
Project Page: https://t.co/0U3UV2DMwR
Poster: https://t.co/Idya9NJbkc

[Matthias Niessner @MattNiessner]

📢📢 𝐏𝐫𝐄𝐝𝐢𝐭𝐨𝐫𝟑𝐃: 𝐅𝐚𝐬𝐭 𝐚𝐧𝐝 𝐏𝐫𝐞𝐜𝐢𝐬𝐞 𝟑𝐃 𝐒𝐡𝐚𝐩𝐞 𝐄𝐝𝐢𝐭𝐢𝐧𝐠 📢📢

We propose a training-free 3D shape editing approach that rapidly and precisely edits the regions intended by the user and keeps the rest as is.

Using a quickly brushed mask and a text prompt, we first apply multi-view editing in the 2D domain and then run our merging algorithm in the 3D feature space to ensure that the edited shape is loyal to the input shape.

Project Page: https://t.co/QRRcF1AP7Q
Video: https://t.co/pMOsbpYUKf

Great work by @ErkocZiya @cangumeli Chaoyang Wang @angelaqdai @peter_wonka @hyjameslee @PeiyeZ

[Georg Hess @GeorgHess3]

@jbohnslav Code will be released :)

[Georg Hess @GeorgHess3]

@anasirisarvi @janusch_patas Projection will only allow you to replay exactly the same viewpoints from the collected data. Here we can change both the ego-vehicles location and that of other actors

[Georg Hess @GeorgHess3]

@kylecoolky @janusch_patas Partly to improve the reconstruction quality, partly we want to enable sensor-complete simulation for AD platforms using LiDAR

[Georg Hess @GeorgHess3]

Presenting NeuRAD: Neural Rendering for Autonomous Driving today, 17:15 at #CVPR2024

Come to poster 28 to learn more about our open source, state-of-the-art rendering method, and to grab a NeuRAD sticker!

https://t.co/zFjSQfIeUw

[William Ljungbergh @wljungbergh]

NeuroNCAP: Photorealistic Closed-loop Safety Testing for Autonomous Driving

We leverage neural rendering to create safety-critical AD scenarios and show that SOTA AD models fail drastically.

page: https://t.co/dwXpQHXS8i
code: https://t.co/0PnpmQznoq
hf: https://t.co/2q9Uy73aRv

[Georg Hess @GeorgHess3]

NeuRAD code is live! Check out for our nerfstudio-integrated SOTA neural rendering method for autonomous driving

https://t.co/zFjSQfHH4Y

[Zhenjun Zhao @zhenjun_zhao]

NeuRAD: Neural Rendering for Autonomous Driving

Adam Tonderski, Carl Lindström, Georg Hess, William Ljungbergh, Lennart Svensson, Christoffer Petersson

tl;dr: handle lidar and camera data in 360◦ and decompose the world into static and dynamic elements

https://t.co/OOs30ibjCN