This paper introduces a novel method for real-time radiance field rendering using 3D Gaussian splatting. The method achieves state-of-the-art visual quality and real-time rendering at 1080p resolution, with training times competitive with the fastest previous methods. The key innovations include a 3D Gaussian scene representation that preserves desirable properties of continuous volumetric radiance fields, an interleaved optimization/density control approach for 3D Gaussians, and a fast visibility-aware rendering algorithm that supports anisotropic splatting and enables real-time rendering. The method is evaluated on several established datasets and demonstrates superior performance compared to previous methods in terms of both visual quality and rendering speed. The 3D Gaussian representation allows for efficient optimization and rendering, while the visibility-aware rendering algorithm ensures accurate and fast backpropagation. The method is able to handle complex scenes with large depth complexity, including both indoor and outdoor environments. The results show that the method achieves high-quality real-time rendering with training times that are significantly faster than previous methods. The method is also able to handle synthetic scenes and demonstrates superior performance in terms of PSNR scores. The method is evaluated on a variety of datasets and shows that it can achieve state-of-the-art results in terms of both visual quality and rendering speed. The method is able to handle scenes with complex geometry and large depth complexity, and it is able to achieve high-quality rendering even in challenging conditions. The method is also able to handle scenes with large depth complexity and is able to achieve high-quality rendering in such cases. The method is able to handle scenes with complex geometry and large depth complexity, and it is able to achieve high-quality rendering even in challenging conditions. The method is able to handle scenes with complex geometry and large depth complexity, and it is able to achieve high-quality rendering even in challenging conditions.This paper introduces a novel method for real-time radiance field rendering using 3D Gaussian splatting. The method achieves state-of-the-art visual quality and real-time rendering at 1080p resolution, with training times competitive with the fastest previous methods. The key innovations include a 3D Gaussian scene representation that preserves desirable properties of continuous volumetric radiance fields, an interleaved optimization/density control approach for 3D Gaussians, and a fast visibility-aware rendering algorithm that supports anisotropic splatting and enables real-time rendering. The method is evaluated on several established datasets and demonstrates superior performance compared to previous methods in terms of both visual quality and rendering speed. The 3D Gaussian representation allows for efficient optimization and rendering, while the visibility-aware rendering algorithm ensures accurate and fast backpropagation. The method is able to handle complex scenes with large depth complexity, including both indoor and outdoor environments. The results show that the method achieves high-quality real-time rendering with training times that are significantly faster than previous methods. The method is also able to handle synthetic scenes and demonstrates superior performance in terms of PSNR scores. The method is evaluated on a variety of datasets and shows that it can achieve state-of-the-art results in terms of both visual quality and rendering speed. The method is able to handle scenes with complex geometry and large depth complexity, and it is able to achieve high-quality rendering even in challenging conditions. The method is also able to handle scenes with large depth complexity and is able to achieve high-quality rendering in such cases. The method is able to handle scenes with complex geometry and large depth complexity, and it is able to achieve high-quality rendering even in challenging conditions. The method is able to handle scenes with complex geometry and large depth complexity, and it is able to achieve high-quality rendering even in challenging conditions.