Search papers, labs, and topics across Lattice.
ImprovedGS+ is a C++/CUDA reimplementation of the ImprovedGS 3D Gaussian Splatting strategy, designed for high performance and low latency within the LichtFeld-Studio framework. It introduces a Long-Axis-Split (LAS) CUDA kernel, custom Laplacian-based importance kernels with NMS, and an adaptive Exponential Scale Scheduler to minimize host-device synchronization. Experiments on Mip-NeRF360 show that ImprovedGS+ achieves a new Pareto-optimal front, outperforming state-of-the-art baselines with faster training, fewer Gaussians, and improved visual quality.
By ditching Python for optimized C++/CUDA kernels, ImprovedGS+ slashes 3D Gaussian Splatting training time by 26.8% while using 13.3% fewer Gaussians and maintaining superior visual quality.
Recent advancements in 3D Gaussian Splatting (3DGS) have shifted the focus toward balancing reconstruction fidelity with computational efficiency. In this work, we propose ImprovedGS+, a high-performance, low-level reinvention of the ImprovedGS strategy, implemented natively within the LichtFeld-Studio framework. By transitioning from high-level Python logic to hardware-optimized C++/CUDA kernels, we achieve a significant reduction in host-device synchronization and training latency. Our implementation introduces a Long-Axis-Split (LAS) CUDA kernel, custom Laplacian-based importance kernels with Non-Maximum Suppression (NMS) for edge scores, and an adaptive Exponential Scale Scheduler. Experimental results on the Mip-NeRF360 dataset demonstrate that ImprovedGS+ establishes a new Pareto-optimal front for scene reconstruction. Our 1M-budget variant outperforms the state-of-the-art MCMC baseline by achieving a 26.8% reduction in training time (saving 17 minutes per session) and utilizing 13.3% fewer Gaussians while maintaining superior visual quality. Furthermore, our full variant demonstrates a 1.28 dB PSNR increase over the ADC baseline with a 38.4% reduction in parametric complexity. These results validate ImprovedGS+ as a scalable, high-speed solution that upholds the core pillars of Speed, Quality, and Usability within the LichtFeld-Studio ecosystem.
