The field of 3D Gaussian Splatting is moving towards more efficient and accurate scene representation and reconstruction. Researchers are exploring new methods to improve the distribution of Gaussians, reduce the number of primitives required, and enhance the registration and fusion of multiple 3D-GS sub-maps. Notable advancements include the development of per-Gaussian optimization techniques, neural shell texture splatting, and automated registration and fusion methods. These innovations have the potential to significantly improve the quality and efficiency of 3D scene representation and reconstruction. Noteworthy papers include: Gaussian Set Surface Reconstruction through Per-Gaussian Optimization, which proposes a method to distribute Gaussians evenly along the latent surface. Fast Learning of Non-Cooperative Spacecraft 3D Models through Primitive Initialization, which contributes a CNN-based primitive initializer for 3DGS using monocular images. Neural Shell Texture Splatting: More Details and Fewer Primitives, which introduces a neural shell texture to disentangle geometry and appearance in Gaussian Splatting. Automated 3D-GS Registration and Fusion via Skeleton Alignment and Gaussian-Adaptive Features, which presents a novel approach for automated 3D-GS sub-map alignment and fusion. UFV-Splatter: Pose-Free Feed-Forward 3D Gaussian Splatting Adapted to Unfavorable Views, which introduces a pose-free, feed-forward 3DGS framework designed to handle unfavorable input views.