cs.CVSep 28, 2026
SaveRRG-SLAM: Real-time Reflection-aware Gaussian SLAM for Indoor Scenes
Organizations: State Key Laboratory of CAD&CG, Zhejiang University, China
Abstract
We introduce the first real-time reflection-aware Gaussian SLAM system for indoor scenes. The system features a reflection-aware TSDF-Gaussian hybrid representation that explicitly separates diffuse scene appearance from reflection components. The base scene is modeled by a TSDF volume and a set of base Gaussians capturing geometry and diffuse appearance, while planar reflections are represented by reflection Gaussian groups associated with detected reflective planes. The rendering is performed in three passes: TSDF raycasting first yields surface color, depth, plane IDs and reflection masks; base Gaussians are then rendered order-independently with depth culling and combined with the TSDF output to form the base image; finally, under the guidance of the plane ID map, reflection Gaussians from different reflection groups are rasterized only into their corresponding planar regions to generate the reflection image, which is subsequently composited with the base image via the reflection mask to produce the final output. For online reconstruction, our system first estimates the camera pose through reflection-aware tracking to suppress interference of reflection-dominated regions. It then identifies reflective planes using geometric, semantic, and temporal cues, and fuses the observations into the augmented TSDF volume with reflection-aware attributes. Afterwards the base and reflection Gaussians are initialized, optimized, and pruned online to maintain both reconstruction quality and efficiency. Experiments on a variety of datasets show that our method outperforms existing SLAM systems in reconstruction quality, tracking robustness, and novel-view rendering for indoor environments with reflections, while preserving real-time performance.
Figures & tables
Figure 1. A meeting room containing strong planar reflections from a screen and polished tables. For clearer visualization, we crop out the ceiling and glass doors. We compare our method with state-of-the-art Gaussian SLAM approaches (GPS-SLAM ( Peng et al., 2025 ) , RTG-SLAM ( Peng et al., 2024 ) , and GS-ICP SLAM ( Ha et al., 2024 ) ) for mesh reconstruction and novel-view synthesis. Regions with large drifting errors are highlighted with red boxes. Our method enables more stable tracking and improves rendering quality in reflective environments. Comparison of Gaussian SLAM methods in a reflective indoor scene with reconstructed geometry and rendered novel views.
Figure 2. Overview of our reflection-aware scene representation and three-pass rendering pipeline. The representation decomposes the scene into a base scene (TSDF volume and base Gaussians) capturing geometry and diffuse appearance, and reflection groups modeling planar reflections. Rendering proceeds in three passes: (1) TSDF raycasting produces color, depth, plane ID, and reflection mask; (2) base Gaussians are rendered with depth culling and combined with TSDF color to produce the base color; (3) reflection Gaussians are rendered conditioned on their associated plane IDs and composited using the reflection mask to generate the final image. This decomposition enables explicit modeling of reflective content while maintaining efficient rendering of the base scene. A schematic overview of the proposed reflection-aware scene representation and rendering pipeline. The figure shows a base scene represented by a TSDF volume and base Gaussians, together with plane-associated reflection groups. The rendering pipeline consists of TSDF raycasting, base Gaussian rendering, and reflection Gaussian rendering, which are composited to produce the final image.
Figure 3. Overview of the proposed reflection-aware RGB-D SLAM pipeline. Given an RGB-D stream, the system performs reflection-aware tracking, reflective plane identification, and online map reconstruction. The tracking module first estimates an initial pose using ICP on depth-derived 3D points, and then refines the pose using reliable color features selected by the rendered tracking mask. The reflective plane identification module combines geometric plane segmentation, semantic segmentation, and TSDF color variance to detect global reflective planes and generate a predicted reflection mask. Finally, the online map reconstruction performs augmented TSDF fusion and Gaussian optimization. A pipeline diagram of a reflection-aware RGB-D SLAM system. The left side shows an RGB-D input stream. The top middle block illustrates reflection-aware tracking, including depth back-projection, ICP-based initial pose estimation, reflection-mask rendering, reliable feature extraction, and feature-based pose refinement. The bottom middle block shows reflective plane identification, where CAPE plane segmentation, GSAM2 semantic segmentation, and TSDF color variance are combined through geometric-semantic association and temporal verification to obtain global reflective planes and a predicted reflection mask. The right block shows hybrid map reconstruction, where color, depth, and reflection-mask information are fused into an augmented TSDF volume and used for Gaussian optimization.
Figure 4. Qualitative comparison of rendering quality across multiple reflective indoor scenes. The first two columns show the original GT images and the cropped regions for clearer visualization of reflective areas. Our method consistently produces sharper and more stable reflections. Red boxes indicate regions where baseline methods (GPS-SLAM, GauS-SLAM, GS-ICP SLAM, RTG-SLAM, and SplaTAM) suffer from drifting or fail to reconstruct reflections accurately. Note that SplaTAM runs out of memory (OOM) on Meeting Room , so no result is reported for it. A qualitative comparison figure of rendering quality results on reflective indoor scenes. Each row shows one scene with the ground-truth image, a cropped reflective region, and rendering results from several baseline methods and the proposed method. Red boxes highlight regions with reflection-related artifacts, such as blurry reflections, noisy reflective surfaces, drift, or missing reflective content. The proposed method produces sharper and more stable reflections compared with the baselines.
Figure 5. Visualization of rendering components on four representative scenes. From left to right, we show the base color image, reflection mask, weighted reflection color image, final rendered image, and ground-truth image. The base color mainly captures diffuse appearance, while the reflection mask localizes reflective regions and the weighted reflection color image shows the contribution of reflection Gaussians. The final image combines the base and reflection components to better reproduce reflective appearance. A visualization figure showing rendering decomposition results on four scenes. Each row corresponds to one scene, and the columns show the base color image, reflection mask, weighted reflection color image, final rendered image, and ground-truth image. The figure illustrates how the proposed method separates diffuse base appearance from reflective components and combines them to produce the final rendering.
Figure 6. Comparison of reconstructed meshes in reflective indoor scenes ( Meeting Room and Private Room ). Our method produces more complete and accurate reconstructions in regions containing a glossy tiled floor and a TV screen. Baseline methods, including GPS-SLAM ( Peng et al., 2025 ) , RTG-SLAM ( Peng et al., 2024 ) , GS-ICP SLAM ( Ha et al., 2024 ) , GauS-SLAM ( Su et al., 2025 ) , BundleFusion ( Dai et al., 2017 ) , and ORB-SLAM2 ( Mur-Artal and Tardós, 2017 ) , exhibit drifting or misalignment in these reflective regions. A qualitative comparison figure of reconstructed meshes in two reflective indoor scenes, Meeting Room and Private Room. The figure compares the proposed method with several baseline SLAM methods. Reflective regions such as glossy tiled floors and TV screens are highlighted. The baseline methods show artifacts such as drift, misalignment, incomplete geometry, or blurred reconstruction, while the proposed method produces cleaner and more consistent mesh reconstruction in reflective regions.
| Method | Metric | RIRD-Syn | ScanNet++ | RIRD-Real | Replica |
| SplaTAM | PSNR | 18.66 | 23.85 | 22.69 | 34.14 |
| SSIM | 0.634 | 0.834 | 0.820 | 0.936 | |
| LPIPS | 0.422 | 0.241 | 0.355 | 0.142 | |
| RTG-SLAM | PSNR | 17.81 | 22.37 | 25.16 | 34.04 |
| SSIM | 0.650 | 0.783 | 0.855 | 0.924 | |
| LPIPS | 0.444 | 0.299 | 0.359 | 0.183 |
Table 1. Rendering quality comparison on RIRD-Syn, ScanNet++, RIRD-Real, and Replica. We report PSNR, SSIM, and LPIPS. RIRD-Syn and ScanNet++ are evaluated under the novel-view setting, while RIRD-Real and Replica are evaluated on training views. RIRD-Syn, ScanNet++, and RIRD-Real contain reflective indoor scenes, whereas Replica is used to evaluate performance on general indoor scenes without reflections. A quantitative table comparing rendering quality across four datasets: RIRD-Syn, ScanNet++, RIRD-Real, and Replica. The rows list different SLAM methods, and each method is evaluated using PSNR, SSIM, and LPIPS. The proposed method achieves the best performance on the reflective datasets and remains competitive on Replica.
Figure 7. Reconstruction and novel-view synthesis results on six real RIRD-Real scenes. For each scene, we show the reconstructed scene overview and zoomed-in views of reflective regions. Colored boxes link the overview to the corresponding zoomed-in renderings. Our method produces coherent scene-scale reconstruction and visually consistent reflective appearance in real reflective indoor environments. A qualitative figure showing reconstruction and novel-view synthesis results on six real indoor scenes from RIRD-Real, including Meeting Room, Rest Room, Living Room 0, Living Room 1, Private Room, and Kitchen. Each scene contains a reconstructed scene overview and two zoomed-in views of reflection-rich regions. Colored boxes link regions in the reconstructed overview to the corresponding zoomed-in renderings. The figure illustrates that the proposed method reconstructs complete scene layouts and preserves reflective appearance on surfaces such as floors, tables, TV screens, and wall paintings.
| Method | RIRD-Syn | Replica | TUM |
| ORB-SLAM2 | 22.26 | 0.36 | 1.27 |
| BundleFusion | 8.52 | 0.42 | 1.53 |
| SplaTAM | 36.15 | 0.39 | 1.81 |
| RTG-SLAM | 10.71 | 0.19 | 1.13 |
| GauS-SLAM | 41.81 | 0.07 | 1.42 |
| GS-ICP SLAM | 16.62 | 0.18 | 2.33 |
Table 2. Tracking accuracy comparison on RIRD-Syn, Replica, and TUM-RGBD. We report the absolute trajectory error (ATE), where lower values indicate better tracking accuracy. RIRD-Syn contains reflective synthetic scenes, while Replica and TUM-RGBD are used to evaluate tracking performance on general indoor scenes. A quantitative table comparing tracking accuracy across RIRD-Syn, Replica, and TUM-RGBD. The rows list different SLAM methods, and the columns report absolute trajectory error for each dataset. The proposed method achieves the lowest error on RIRD-Syn and TUM-RGBD, and competitive performance on Replica.
Figure 8. Visualization of reflective plane identification. From left to right, we show the input RGB-D frame, plane segmentation map, semantic segmentation map, TSDF variance map, and predicted reflection mask. Different colors in the plane segmentation map indicate different planar regions. In the semantic segmentation map, TV , table , and floor indicate the semantic categories. The TSDF variance map is colorized according to temporal color variance, where brighter colors indicate higher variance. By combining geometric, semantic, and temporal cues, our method predicts reliable reflective regions for reflection-aware reconstruction. An overview figure showing the reflective plane identification process. The figure contains five columns: an input RGB-D frame, a plane segmentation map with different planar regions shown in different colors, a semantic segmentation map highlighting reflection-prone categories such as TV, table, and floor, a TSDF temporal color variance map where brighter regions indicate higher variance, and the final predicted reflection mask. The figure illustrates how geometric plane segmentation, semantic cues, and temporal variance are combined to identify reflective regions.
| Method | Dataset | Acc | Acc Ratio | Com | Com Ratio |
| SplaTAM | ScanNet++ | 3.10 | 65.55 | 3.03 | 61.21 |
| RIRD-Syn | 6.32 | 44.64 | 4.74 | 50.52 | |
| RTG-SLAM | ScanNet++ | 1.41 | 91.84 | 1.74 | 87.18 |
| RIRD-Syn | 7.85 | 62.69 | 5.77 | 59.09 | |
| GauS-SLAM | ScanNet++ | 3.92 | 74.13 | 1.49 | 95.84 |
| RIRD-Syn | 60.01 | 63.22 | 45.58 | 64.22 |
Table 3. Geometry reconstruction quality comparison on RIRD-Syn and ScanNet++. We report accuracy, completion, accuracy ratio, and completion ratio. The ratio metrics measure the percentage of points with errors below 3 cm. A quantitative table comparing geometry reconstruction quality on RIRD-Syn and ScanNet++. The rows list different SLAM methods, and each method is evaluated using accuracy, accuracy ratio, completion, and completion ratio. The proposed method achieves the best geometry quality on RIRD-Syn and competitive results on ScanNet++.
| Method | Dataset | Mapping /Frame (ms) | FPS | Gaussian Count | Memory (MB) | PSNR |
| SplaTAM | Replica | 3066.7 | 0.3 | 6.22M | 9551 | 38.18 |
| RIRD-Real | – | – | – | OOM | – | |
| RTG-SLAM | Replica | 59.9 | 16.6 | 0.82M | 2755 | 37.78 |
| RIRD-Real | 82.4 | 12.12 | 1M | 10656 | 22.42 | |
| GauS-SLAM | Replica | 314.9 | 3.1 | 6.69M | 15745 | 42.40 |
| RIRD-Real | 376.1 | 2.6 | 10.05M | 23891 | 17.89 |
Table 4. Runtime, memory, and rendering quality comparison on Replica Office 0 and RIRD-Real Meeting Room . We report mapping time per frame, FPS, Gaussian count, GPU memory usage, and PSNR. Replica Office 0 represents a general indoor scene, while Meeting Room contains strong planar reflections. A quantitative table comparing runtime, memory usage, Gaussian count, and PSNR across different SLAM methods on Replica office 0 and the RIRD-Real Meeting Room scene. The proposed method achieves high runtime and compact Gaussian count, especially on the reflective Meeting Room scene, while maintaining the best PSNR among the compared methods.
| Module | Time / Frame (ms) |
| Pose Tracking | 0.85 |
| Plane Detection | 1.88 |
| Semantic Segmentation | 7.04 |
| TSDF Fusion | 0.19 |
| Pre-mapping Stage | 10.12 |
| Gaussian Mapping | 14.18 |
Table 5. Runtime breakdown on the RIRD-Real Meeting Room scene. We report the average per-frame runtime of pose tracking, plane detection, semantic segmentation, TSDF fusion, the total pre-mapping stage, and Gaussian mapping. A table reporting the per-frame runtime of major components on the RIRD-Real Meeting Room scene. Pose tracking takes 0.85 ms, plane detection takes 1.88 ms, semantic segmentation takes 7.04 ms, TSDF fusion takes 0.19 ms, the total pre-mapping stage including additional lightweight overhead takes 10.12 ms, and Gaussian mapping takes 14.18 ms.
| Rendering Stage | Time / Frame (ms) | FPS | Num. of Gaussians |
| TSDF Raycasting | 0.72 | 1389.83 | – |
| Base Gaussian Rendering | 1.05 | 950.00 | 0.30M |
| Reflection Gaussian Rendering | 1.39 | 716.19 | 0.34M |
| Total | 3.22 | 310.00 | 0.64M |
Table 6. Rendering performance breakdown on the RIRD-Real Meeting Room scene. We report the per-frame runtime, FPS, and number of Gaussians for each rendering stage. A table reporting the rendering speed of the proposed system on the RIRD-Real Meeting Room scene. The rows include TSDF raycasting, base Gaussian rendering, reflection Gaussian rendering, and total rendering. The table reports runtime per frame, FPS, and the number of Gaussians used in each Gaussian rendering stage.
Figure 9. Ablation study of plane-conditioned reflection rendering on the Room 2 scene from RIRD-Syn. Without plane grouping, reflections become dim and blurry. Plane-associated reflection groups produce more accurate reflected content, while our single-pass plane-filtered rendering achieves comparable quality to multi-pass masked compositing with higher efficiency. A qualitative ablation figure comparing reflection rendering results on a synthetic indoor scene. The figure shows results from variants without plane grouping, without single-pass filtering, and the full model. The variant without plane grouping produces weaker and blurrier reflections, while the plane-conditioned variants produce clearer and more accurate reflected content. The full model preserves similar reflection quality to the multi-pass variant while using a more efficient single-pass rendering strategy.
| Ablation | Base Render (ms) | Refl. Render (ms) | Render FPS | PSNR |
| w/o Plane Grouping | 3.37 | 0.33 | 265.4 | 26.55 |
| w/o Single-Pass Filtering | 3.32 | 1.39 | 209.2 | 26.90 |
| Full Model | 3.36 | 0.29 | 268.8 | 27.09 |
Table 7. Ablation study of reflection grouping and rendering on RIRD-Syn Room 2 . We report base rendering time, reflection rendering time, total rendering FPS, and PSNR. The results show that plane grouping improves reflection quality, while our single-pass filtering maintains high rendering efficiency. A table reporting the ablation study of reflection grouping and rendering. The compared variants include removing plane grouping, removing single-pass filtering, and the full model. The table reports base rendering time, reflection rendering time, rendering FPS, and PSNR. The full model achieves the best PSNR and rendering FPS while keeping reflection rendering efficient.
Figure 10. Ablation study of reflection-aware tracking. We compare three variants: w/o ORB refinement , which uses ICP only; w/o tracking mask , which applies ORB refinement without masking reflection-dominated regions; and the Full model , which uses masked ORB refinement. The results show that unmasked ORB refinement can introduce drift in reflective regions, while the full model produces cleaner reconstruction. A qualitative ablation figure for reflection-aware tracking. The figure compares reconstructed meshes under three variants: ICP-only tracking without ORB refinement, ORB refinement without reflection masking, and the full model with reflection-aware masked ORB refinement. The variant without reflection masking shows visible drift in reflective regions, while the full model produces a cleaner and more stable reconstruction.
| Ablation | ATE | Acc. | Acc. Ratio | Com. | Com. Ratio |
| w/o ORB Refinement | 0.56 | 0.95 | 97.54 | 0.96 | 98.64 |
| w/o Tracking Mask | 1.21 | 3.19 | 86.72 | 0.93 | 92.52 |
| Full Model | 0.31 | 0.93 | 98.86 | 0.81 | 99.82 |
Table 8. Quantitative ablation study of reflection-aware tracking on RIRD-Syn Room 1 . We report ATE for tracking accuracy and geometry metrics including accuracy, completion, and their corresponding ratios. A quantitative ablation table evaluating reflection-aware tracking. The compared variants include removing ORB refinement, removing the reflection mask, and the full model. The table reports ATE, geometry accuracy, accuracy ratio, completion, and completion ratio. The full model achieves the best tracking accuracy and geometry reconstruction quality.
Figure 11. Ablation study of the reflection color regularization term on RIRD-Real Living Room 0 . We compare the results without and with . From left to right, we show the base color image, weighted reflection color image, final composited image, and color error map. With , the base and reflection components are better constrained during optimization, resulting in lower color errors and improved final rendering quality. A qualitative ablation figure comparing the method without and with the reflection color regularization term on the RIRD-Real Living Room 0 scene. The top row shows results without the regularization term, and the bottom row shows results with the regularization term. The columns show the base color image, weighted reflection color image, final composited image, and color error map. Without the regularization term, the base and reflection components exhibit stronger interference and the color error map shows larger errors. With the regularization term, the color error is reduced and the final rendering quality is improved.
| Ablation | PSNR | SSIM | LPIPS |
| w/o | 28.36 | 0.888 | 0.231 |
| w/ | 30.38 | 0.919 | 0.224 |
Table 9. Ablation study of the reflection color regularization term on RIRD-Real Living Room 0 . We compare the rendering quality with and without . A table comparing the effect of the reflection color regularization term on the Living Room 0 scene from RIRD-Real. Without the regularization term, the method achieves 28.36 PSNR, 0.888 SSIM, and 0.231 LPIPS. With the regularization term, the method improves to 30.38 PSNR, 0.919 SSIM, and 0.224 LPIPS.
| Threshold | PSNR | SSIM | LPIPS |
| 26.23 | 0.872 | 0.225 | |
| =0.005 | 26.37 | 0.881 | 0.214 |
| =0.01 | 26.47 | 0.883 | 0.211 |
| =0.03 | 26.27 | 0.879 | 0.209 |
| =0.05 | 25.65 | 0.873 | 0.253 |
| =0.1 | 24.06 | 0.864 | 0.243 |
Table 10. Ablation study of the temporal color variance threshold on RIRD-Syn Room 2 . We report PSNR, SSIM, and LPIPS under different threshold values to evaluate the sensitivity of reflective-region verification. A quantitative ablation table evaluating the temporal color variance threshold used for reflective-region verification. The rows correspond to different values of delta v, and the columns report PSNR, SSIM, and LPIPS. The results show that the method remains stable across a moderate range of threshold values.
Appendix figures & tables10 assets
Supplementary material from the paper’s appendix.
Appendix
| Statistic | LR0 | LR1 | Rest | Meeting | Private | Kitchen |
| Frames | 1570 | 2080 | 3470 | 3699 | 1649 | 2670 |
| Traj. Len. (m) | 15.3 | 20.1 | 33.7 | 59.1 | 20.7 | 26.4 |
| Area (m 2 ) | 19.2 | 15.2 | 41.7 | 102.2 | 21.6 | 36.7 |
Table 1. Statistics of the RIRD-Real dataset. LR0 and LR1 denote the two living-room scenes. We report the number of frames, trajectory length, and scanned area for each scene. A table summarizing the RIRD-Real dataset. The columns correspond to six real indoor scenes, including two living rooms, a restroom, a meeting room, a private room, and a kitchen. The rows report the number of frames, trajectory length in meters, and scanned area in square meters for each scene.
| Statistic | Room 0 | Room 1 | Room 2 |
| Frames | 1681 | 2337 | 2206 |
| Traj. Len. (m) | 41.9 | 58.5 | 50.6 |
| Area (m 2 ) | 43.8 | 43.25 | 97.8 |
Table 2. Statistics of the RIRD-Syn dataset. We report the number of frames, trajectory length, and scanned area for each synthetic scene. A table summarizing the RIRD-Syn dataset. The columns correspond to three synthetic indoor scenes, Room 0, Room 1, and Room 2. The rows report the number of frames, trajectory length in meters, and scanned area in square meters for each scene.
| Method | Room 0 | Room 1 | Room 2 | Average |
| ORB-SLAM2 | 63.45 | 1.35 | 1.98 | 22.26 |
| BundleFusion | 22.40 | 0.92 | 2.23 | 8.52 |
| SplaTAM | 103.32 | 3.00 | 2.13 | 36.15 |
| RTG-SLAM | 30.24 | 0.43 | 1.46 | 10.71 |
| GauS-SLAM | 116.76 | 6.64 | 2.03 | 41.81 |
| GS-ICP SLAM | 25.00 | 0.29 | 24.56 | 16.62 |
Table 3. Per-scene tracking accuracy comparison on RIRD-Syn. We report ATE for each synthetic reflective scene and the average across all three scenes. A quantitative table comparing tracking accuracy on the RIRD-Syn dataset. The rows list different SLAM methods, and the columns report ATE on Room 0, Room 1, and Room 2, together with the average performance. The proposed method achieves the lowest ATE on all three synthetic reflective scenes and the best average tracking accuracy.
| Method | fr1_desk | fr2_xyz | fr3_office | Average |
| ORB-SLAM2 | 1.62 | 0.45 | 1.75 | 1.27 |
| BundleFusion | 1.65 | 1.13 | 1.81 | 1.53 |
| SplaTAM | 1.62 | 1.30 | 2.53 | 1.81 |
| RTG-SLAM | 1.51 | 0.58 | 1.32 | 1.13 |
| GauS-SLAM | 1.48 | 1.29 | 1.50 | 1.42 |
| GS-ICP SLAM | 2.96 | 1.77 | 2.26 | 2.33 |
Table 4. Per-sequence tracking accuracy comparison on TUM-RGBD. We report ATE for each evaluated sequence and the average, where lower values indicate better tracking accuracy. A quantitative table comparing tracking accuracy on the TUM-RGBD dataset. The rows list different SLAM methods, and the columns report ATE on fr1_desk, fr2_xyz, and fr3_office, together with the average performance. The proposed method achieves the lowest average ATE among the compared methods.
| Method | Rm 0 | Rm 1 | Rm 2 | Off 0 | Off 1 | Off 2 | Off 3 | Off 4 | Avg. |
| ORB-SLAM2 | 0.34 | 0.38 | 0.38 | 0.33 | 0.32 | 0.41 | 0.34 | 0.37 | 0.36 |
| BundleFusion | 0.41 | 0.40 | 0.43 | 0.43 | 0.37 | 0.42 | 0.46 | 0.44 | 0.42 |
| SplaTAM | 0.38 | 0.23 | 0.28 | 0.48 | 0.32 | 0.38 | 0.43 | 0.64 | 0.39 |
| RTG-SLAM | 0.20 | 0.19 | 0.12 | 0.16 | 0.13 | 0.22 | 0.25 | 0.25 | 0.19 |
| GauS-SLAM | 0.06 | 0.08 | 0.07 | 0.06 | 0.04 | 0.09 | 0.07 | 0.06 | 0.07 |
| GS-ICP SLAM | 0.16 | 0.17 | 0.12 | 0.19 | 0.13 | 0.17 | 0.18 | 0.23 | 0.18 |
Table 5. Per-scene tracking accuracy comparison on Replica. We report ATE for each evaluated room and office scene, together with the average across all scenes. A quantitative table comparing tracking accuracy on the Replica dataset. The rows list different SLAM methods, and the columns report ATE on three room scenes and five office scenes, together with the average performance. The proposed method achieves competitive tracking accuracy on this general indoor benchmark.
Figure 1. Tracking-error and mesh visualization on RIRD-Syn. We compare our method with GPS-SLAM, GS-ICP SLAM, and GauS-SLAM on three synthetic reflective scenes. The camera trajectories are colored by frame-wise tracking error, where green indicates lower error and red indicates larger error. Our method maintains stable trajectories and produces more coherent mesh reconstruction in reflective scenes. A qualitative comparison figure showing reconstructed meshes and tracking-error-colored camera trajectories on three RIRD-Syn scenes: Room 0, Room 1, and Room 2. The columns compare Ours, GPS-SLAM, GS-ICP SLAM, and GauS-SLAM. A color bar indicates tracking error from green to red, corresponding to low to high error. The proposed method shows more stable trajectories and cleaner reconstructed meshes, while some baseline methods exhibit larger tracking errors and visible reconstruction distortions.
| Method | Metric | a5859cfd40 | 8e6ff28354 | Average |
| SplaTAM | PSNR | 21.08 | 26.63 | 23.85 |
| SSIM | 0.781 | 0.887 | 0.834 | |
| LPIPS | 0.293 | 0.190 | 0.241 | |
| RTG-SLAM | PSNR | 18.67 | 26.07 | 22.37 |
| SSIM | 0.705 | 0.862 | 0.783 | |
| LPIPS | 0.372 | 0.226 | 0.299 |
Table 6. Per-scene novel-view rendering quality comparison on the selected reflective ScanNet++ scenes. We report PSNR, SSIM, and LPIPS for each scene and the average across the two scenes. A quantitative table comparing novel-view rendering quality on two selected reflective ScanNet++ scenes. The rows list different SLAM methods, and each method is evaluated using PSNR, SSIM, and LPIPS on scenes a5859cfd40 and 8e6ff28354, together with the average performance. The proposed method achieves the best average rendering quality across the reported metrics.
| Method | Metric | Room 0 | Room 1 | Room 2 | Average |
| SplaTAM | PSNR | 15.43 | 23.45 | 17.08 | 18.66 |
| SSIM | 0.592 | 0.764 | 0.547 | 0.634 | |
| LPIPS | 0.474 | 0.341 | 0.453 | 0.422 | |
| RTG-SLAM | PSNR | 14.74 | 22.97 | 15.74 | 17.81 |
| SSIM | 0.605 | 0.813 | 0.533 | 0.650 | |
| LPIPS | 0.496 | 0.332 | 0.503 | 0.444 |
Table 7. Per-scene novel-view rendering quality comparison on RIRD-Syn. We report PSNR, SSIM, and LPIPS for each synthetic reflective scene and the average across all three scenes. A quantitative table comparing novel-view rendering quality on the RIRD-Syn dataset. The rows list different SLAM methods, and each method is evaluated using PSNR, SSIM, and LPIPS on Room 0, Room 1, and Room 2, together with the average performance. The proposed method achieves the best average rendering quality across the reported metrics.
| Method | Metrics | Living Room 0 | Living Room 1 | Kitchen | Private Room | Rest Room | Meeting Room | Average |
| SplaTAM | PSNR | 24.00 | 26.72 | 24.21 | 24.44 | 22.71 | 14.03 | 22.69 |
| SSIM | 0.856 | 0.902 | 0.843 | 0.844 | 0.797 | 0.675 | 0.820 | |
| LPIPS | 0.352 | 0.281 | 0.336 | 0.277 | 0.378 | 0.507 | 0.355 | |
| RTG-SLAM | PSNR | 25.42 | 25.39 | 25.16 | 25.34 | 27.25 | 22.42 | 25.16 |
| SSIM | 0.863 | 0.881 | 0.861 | 0.866 | 0.888 | 0.773 | 0.855 | |
| LPIPS | 0.365 | 0.345 | 0.371 | 0.302 | 0.317 | 0.455 | 0.359 |
Table 8. Per-scene input-view rendering quality comparison on RIRD-Real. We report PSNR, SSIM, and LPIPS for each real scene and the average across all six scenes. A quantitative table comparing input-view rendering quality on the RIRD-Real dataset. The rows list different SLAM methods, and each method is evaluated using PSNR, SSIM, and LPIPS on six real reflective indoor scenes: Living Room 0, Living Room 1, Kitchen, Private Room, Rest Room, and Meeting Room, together with the average performance. The proposed method achieves the best average rendering quality across all reported metrics.
| Method | Metric | Office 0 | Office 1 | Office 2 | Office 3 | Office 4 | Room 0 | Room 1 | Room 2 | Average |
| SplaTAM | PSNR | 38.18 | 39.26 | 32.00 | 30.41 | 32.41 | 32.36 | 33.45 | 35.01 | 34.14 |
| SSIM | 0.963 | 0.958 | 0.929 | 0.903 | 0.916 | 0.933 | 0.930 | 0.952 | 0.936 | |
| LPIPS | 0.118 | 0.139 | 0.158 | 0.166 | 0.179 | 0.118 | 0.133 | 0.126 | 0.142 | |
| RTG-SLAM | PSNR | 37.78 | 38.41 | 32.23 | 31.86 | 34.98 | 30.41 | 32.82 | 33.82 | 34.04 |
| SSIM | 0.952 | 0.950 | 0.918 | 0.911 | 0.935 | 0.882 | 0.917 | 0.929 | 0.924 | |
| LPIPS | 0.141 | 0.182 | 0.212 | 0.197 | 0.173 | 0.196 | 0.179 | 0.183 | 0.183 |
Table 9. Per-scene input-view rendering quality comparison on Replica. We report PSNR, SSIM, and LPIPS for each evaluated scene and the average across all scenes. A quantitative table comparing input-view rendering quality on the Replica dataset. The rows list different SLAM methods, and each method is evaluated using PSNR, SSIM, and LPIPS on office and room scenes, together with the average performance. The proposed method remains competitive on this general indoor benchmark.