What is the difference between NeRF and Gaussian Splatting technology?

Daniel Pikl
2 min readSep 4, 2023

NeRF (Neural Radiance Fields) and Gaussian Splatting are two different techniques used in computer graphics and computer vision.

  1. NeRF (Neural Radiance Fields):
  2. . – NeRF is a novel technique that uses deep learning to model 3D scenes from 2D images or image collections.
  3. . – It represents a scene as a continuous 3D function that predicts the color and opacity of a point in space.
  4. . – NeRF aims to capture complex 3D scenes by training neural networks to learn the underlying scene representation from multiple 2D images, allowing for the synthesis of novel views of the scene.
  5. . – It has gained popularity for its ability to generate high-quality 3D reconstructions and novel views of scenes.

2. Gaussian Splatting:

. – Gaussian splatting is a traditional computer graphics technique used for volume rendering and point-based rendering.

. – It involves projecting 3D data, often represented as points with attributes, onto a 2D image plane.

. – Each point in 3D space is „splat“ or projected onto the 2D image plane using a Gaussian distribution, creating a smooth contribution to the image pixel values.

. – Gaussian splatting is typically used for rendering volumetric data, such as medical scans or particle simulations, to generate 2D images that convey the 3D information.

In summary, NeRF is a recent deep learning-based technique for 3D scene reconstruction and novel view synthesis, while Gaussian splatting is a traditional method used for volume rendering and point-based rendering, often applied in contexts like medical imaging and scientific visualization. The choice between them depends on the specific application and requirements of the task at hand.