BakedSDF Announced by Google Research

BakedSDF Announced by Google Research

BakedSDF Announced by Google Research

Michael Rubloff

Michael Rubloff

Mar 4, 2023

Email
Copy Link
Twitter
Linkedin
Reddit
Whatsapp
BakedSDF
BakedSDF

Rounding out the busy month of February for NeRFs, another paper has been announced by Google Research.

This method also allows for NeRFs to be displayed on common hardware such as a web browser or phone with a photo realistic and high resolution mesh. Additionally, BakedSDF allows for a high frame rate to be maintained as a user explores the scene. In the below demo, the scene is demonstrated to achieve frame rates above 60fps, while Instant-NGP is only able to achieve >10 fps. While MobileNeRF is able to render ~50 fps, the underlying mesh is clunky and doesn't represent the scene cleanly.

However, BakedSDF captures the best of both worlds by achieving a high frame rate and representing the underlying mesh cleanly. The method also supports Diffuse and Specular colors, while also allowing for accurately relighting of the scene. While only briefly mentioned, there is also the ability to edit the appearance, which is quite underrated especially as people continue to explore Control Net and Gen-1.

This is yet another contribution to the field, no pun intended, for Google Research Scientist team.

The official video of BakedSDF

I highly suggest trying out the 8 demos for yourself to further envision how the majority of people will be interacting with NeRFs in the near future.

My sources have been telling me that this upcoming week will see updates and announcements that will be considered to be "major" and will catapult NeRF forwards. Stay tuned here for the upcoming news and developments.

While several of the most recent developments have been published without authors, NeRF creator, Matt Tancik explained why below:

Featured

Recents

Featured

Platforms

3DGS to Dense Point Cloud V2

3DGS to Dense Point Cloud returns with awesome updates!

Michael Rubloff

Apr 16, 2025

Platforms

3DGS to Dense Point Cloud V2

3DGS to Dense Point Cloud returns with awesome updates!

Michael Rubloff

Apr 16, 2025

Platforms

3DGS to Dense Point Cloud V2

3DGS to Dense Point Cloud returns with awesome updates!

Michael Rubloff

Platforms

Preferred Networks Inc. Announces 3DGS Unreal Engine 5 Plugin

Tokyo based Preferred Networks Inc has announced an exciting upcoming UE 5 Plugin.

Michael Rubloff

Apr 15, 2025

Platforms

Preferred Networks Inc. Announces 3DGS Unreal Engine 5 Plugin

Tokyo based Preferred Networks Inc has announced an exciting upcoming UE 5 Plugin.

Michael Rubloff

Apr 15, 2025

Platforms

Preferred Networks Inc. Announces 3DGS Unreal Engine 5 Plugin

Tokyo based Preferred Networks Inc has announced an exciting upcoming UE 5 Plugin.

Michael Rubloff

Platforms

404—GEN: Bringing Text to 3D Gaussian Splatting to Unity

A free text to 3D generator in Unity using 3DGS.

Michael Rubloff

Apr 14, 2025

Platforms

404—GEN: Bringing Text to 3D Gaussian Splatting to Unity

A free text to 3D generator in Unity using 3DGS.

Michael Rubloff

Apr 14, 2025

Platforms

404—GEN: Bringing Text to 3D Gaussian Splatting to Unity

A free text to 3D generator in Unity using 3DGS.

Michael Rubloff

Interview

Interview with Rev Lebaredian: Simulation, Robotics, and the Future of Imaging

At NVIDIA’s GTC conference, Rev Lebaredian discusses how simulation can power the next wave of AI, from lifelike 3D reconstruction to industrial scale robotics and physical intelligence.

Michael Rubloff

Apr 14, 2025

Interview

Interview with Rev Lebaredian: Simulation, Robotics, and the Future of Imaging

At NVIDIA’s GTC conference, Rev Lebaredian discusses how simulation can power the next wave of AI, from lifelike 3D reconstruction to industrial scale robotics and physical intelligence.

Michael Rubloff

Apr 14, 2025

Interview

Interview with Rev Lebaredian: Simulation, Robotics, and the Future of Imaging

At NVIDIA’s GTC conference, Rev Lebaredian discusses how simulation can power the next wave of AI, from lifelike 3D reconstruction to industrial scale robotics and physical intelligence.

Michael Rubloff