[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-77283":3},{"id":4,"name":5,"fullName":6,"owner":5,"repo":5,"description":7,"homepage":8,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":14,"stars7d":16,"stars30d":17,"stars90d":15,"forks30d":15,"starsTrendScore":18,"compositeScore":19,"rankGlobal":9,"rankLanguage":9,"license":20,"archived":21,"fork":21,"defaultBranch":22,"hasWiki":21,"hasPages":21,"topics":23,"createdAt":9,"pushedAt":9,"updatedAt":24,"readmeContent":25,"aiSummary":26,"trendingCount":15,"starSnapshotCount":15,"syncStatus":16,"lastSyncTime":27,"discoverSource":28},77283,"RVizSplat","RVizSplat\u002FRVizSplat","Rendering 3D Gaussian Splats in RViz2","",null,"C++",126,7,118,1,0,2,8,3,2.71,"Apache License 2.0",false,"main",[],"2026-06-12 02:03:42","# RVizSplat\n\nRVizSplat is an RViz2 display plugin that provides end-to-end visualization of 3D Gaussian Splats in RViz.\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"images\u002Frviz_splat.jpg\" alt=\"RVizSplat\" width=\"400\">\n\u003C\u002Fp>\n\n### Build status\n\n[![Rolling](https:\u002F\u002Fgithub.com\u002FRVizSplat\u002FRVizSplat\u002Factions\u002Fworkflows\u002Frolling.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002FRVizSplat\u002FRVizSplat\u002Factions\u002Fworkflows\u002Frolling.yml)    \n[![Kilted](https:\u002F\u002Fgithub.com\u002FRVizSplat\u002FRVizSplat\u002Factions\u002Fworkflows\u002Fkilted.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002FRVizSplat\u002FRVizSplat\u002Factions\u002Fworkflows\u002Fkilted.yml)    \n[![Jazzy](https:\u002F\u002Fgithub.com\u002FRVizSplat\u002FRVizSplat\u002Factions\u002Fworkflows\u002Fjazzy.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002FRVizSplat\u002FRVizSplat\u002Factions\u002Fworkflows\u002Fjazzy.yml)\n\n# How to build and run from source\n\n```bash\nmkdir -p ~\u002Fros_ws\u002Fsrc\ncd ~\u002Fros_ws\u002Fsrc\ngit clone https:\u002F\u002Fgithub.com\u002FRVizSplat\u002FRVizSplat.git\ncd ~\u002Fros_ws\nrosdep update\nrosdep install --from-paths src --ignore-src -r -y\ncolcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=Release\nsource install\u002Fsetup.bash\n```\n\n# Installation via apt\n\nThis feature is currently under development (See https:\u002F\u002Fgithub.com\u002Fros\u002Frosdistro\u002Fpull\u002F50909 for details)\n\nAfter sourcing your ROS 2 environment:\n\n```bash\nsudo apt-get install ros-$ROS_DISTRO-gsplat-rviz-plugin ros-$ROS_DISTRO-gsplat-publisher ros-$ROS_DISTRO-gsplat-msgs\n```\n\n# Examples\n\n## Rendering of a LeRobot SO-100 arm with a scene containing 6 million splats\n\n\u003Cp align=\"left\">\n  \u003Cimg src=\"images\u002Fle_robot_6mil.gif\" alt=\"LeRobot SO-100 arm with 6 million splats\" width=\"500\">\n\u003C\u002Fp>\n\n## Rendering of transparent markers along with other gaussians\n\n\u003Cp align=\"left\">\n  \u003Cimg src=\"images\u002Falpha_0_5_marker.gif\" alt=\"Transparent markers rendered with gaussians\" width=\"500\">\n\u003C\u002Fp>\n\n# Using OIT for performance optimization\n\nIf you have a resource constrained CPU and a weaker GPU (just integrated graphics), you might want to consider bypassing sorting entirely. For this use case, we provide OIT based implementations.\n\nTo activate this, follow the \"Advanced\" options in the RViz plugin and select WBOIT.\n\n\u003Cp align=\"left\">\n  \u003Cimg src=\"images\u002Frviz_wboit.png\" alt=\"RViz WBOIT\" width=\"400\">\n\u003C\u002Fp>\n\n# Architecture\n\nComing soon!\n\n# Evaluation\n\nThe `gsplat_plugin_evaluation\u002Feval.py` script computes image quality metrics (PSNR, SSIM, LPIPS) between a ref_folder and an eval_folder.\n\nImages are matched by the trailing 3-digit number in the filename (e.g. `img_001.png` in the ref_folder is paired with `*_001.png` in the eval_folder).\n\n### Usage\n\n```bash\ncd gsplat_plugin_evaluation\npython eval.py \u003Cref_folder> \u003Ceval_folder> [--metrics psnr ssim lpips] [--lpips-net alex|vgg]\n```\n\n| Argument | Description |\n|---|---|\n| `ref_folder` | Folder containing reference (ground-truth) images |\n| `eval_folder` | Folder containing images to evaluate |\n| `--metrics` | Space-separated list of metrics to compute (default: all three) |\n| `--lpips-net` | Backbone network for LPIPS — `alex` (default) or `vgg` |\n\n### Examples\n\nCompute all metrics using the default AlexNet backbone:\n```bash\npython eval.py data\u002Fref data\u002Feval\n```\n\nCompute PSNR and LPIPS with VGG backbone:\n```bash\npython eval.py data\u002Fref data\u002Feval --metrics psnr lpips --lpips-net vgg\n```\n\n### Output\n\nThe script prints a per-image table and a mean row at the bottom:\n\n```\nImage                    PSNR        SSIM       LPIPS\n--------------------------------------------------------\nimg_001.png           32.1500      0.9210      0.0431\nimg_002.png           29.8300      0.8970      0.0612\n--------------------------------------------------------\nMean                  30.9900      0.9090      0.0522\n```","RVizSplat 是一个 RViz2 显示插件，用于在 RViz 中实现 3D 高斯点云的端到端可视化。该项目使用 C++ 编写，支持大规模高斯点云的高效渲染，并通过 OIT 技术优化了在资源受限设备上的性能表现。它不仅能够处理包含数百万个点的数据集，还允许用户调整透明度和其他参数以满足不同的显示需求。RVizSplat 适用于需要进行复杂三维数据可视化分析的场景，如机器人导航、环境建模等。","2026-06-11 03:55:18","CREATED_QUERY"]