ClearGS: Reliability-Aware Gaussian Splatting from Handheld Videos

Xuanzhi Liu; Xinyi Wu; Hang Pan; Wensi Huang; Zhenyao Wu; Ruize Han; Song Wang

cs.CVSep 25, 2026

ClearGS: Reliability-Aware Gaussian Splatting from Handheld Videos

Authors: Xuanzhi Liu, Xinyi Wu, Hang Pan, Wensi Huang, Zhenyao Wu, Ruize Han, Song Wang

Organizations: Shenzhen University of Advanced Technology · HONOR · University of New South Wales · Southern University of Science and Technology

Abstract

We present ClearGS for 3D Gaussian Splatting (3DGS) from handheld videos with uneven viewpoint coverage and mixed frame quality. Rather than selecting frames with binary decisions, ClearGS uses Reliability-aware View Allocation (RVA) to assign graded raw-supervision weights based on appearance reliability, degradation risk, and geometric utility, while weakly reactivating useful suppressed frames to maintain trajectory coverage. Since weighting cannot restore details lost to blur or distortion, ClearGS further introduces Render-Guided In-Video Restoration (RIVR). The current 3DGS render provides a pose-aligned structural candidate, a frozen no-reference restoration expert restores the corresponding raw video observation without any clean reference image, and no-reference perceptual scores select among the render, restored observation, and high-frequency fused candidate. ClearGS then applies Full-Trajectory Repair Consolidation to revisit accepted repairs and preserve details introduced early. On GS2E and GSOTM, ClearGS achieves state-of-the-art overall performance, with consistent CLIP-IQA and MUSIQ gains and LPIPS reductions in most degradation settings, without paired sharp supervision or matched clean references.

Figures & tables

Figure 1: Traditional pipelines either train or select handheld-video frames without graded reliability control, then repair degraded views using nearby references or cross-view cues that may share the same artifacts. ClearGS instead uses RVA to separate appearance reliability from geometric utility, and applies no-reference RIVR to add selected repair evidence without matched clean references.

Figure 2: Overview of ClearGS. RVA assigns frame-wise raw-supervision weights from appearance reliability, degradation risk, and geometric utility, and weakly reactivates useful suppressed frames in uncovered trajectory regions. RIVR selects among the current render, restored raw observation, and detail-fused candidate using no-reference quality scores, then distills accepted pseudo observations with real observations.

Part	Quantity / role	Definition or condition	Parameters / weight
Appearance	$\mathbf{r}_{i}$	ordered quality cues: sharpness, structural clarity, texture reliability, artifact cleanliness, exposure quality	$\mathbf{w}_{R}=(0.30,0.25,0.20,0.15,0.10)$
	$\mathbf{b}_{i}$	ordered degradation cues: motion blur, defocus, ghosting, exposure failure, unreliable texture, artifacts	$\mathbf{w}_{B}=(0.25,0.15,0.20,0.15,0.15,0.10)$
	$R_{i}$	appearance reliability score	$R_{i}=\mathbf{w}_{R}^{\top}\mathbf{r}_{i}$
	$B_{i}$	degradation risk score	$B_{i}=\mathbf{w}_{B}^{\top}\mathbf{b}_{i}$
	$\mathcal{A}$	appearance-reliable frame set	$\tau_{R}=0.70,\ \tau_{B}=0.60$
Geometry	$d_{ij}^{c}$	normalized camera-center distance between frames $i$ and $j$	Euclidean distance normalized to $[0,1]$

Table 1: RVA notation, parameter settings, and supervision policy. The standard supervision conditions are evaluated from top to bottom. Afterward, trajectory coverage compensation may reactivate a suppressed frame in an uncovered trajectory bin with a fixed weak-supervision weight.

Candidate	Construction	Role
Render	$\mathbf{R}_{t}$	Pose-aligned structural candidate
Restored	$\mathbf{D}_{t}$	Restoration of the raw target observation
Fused	$\mathbf{F}_{t}=\mathbf{R}_{t}+\mathbf{D}_{t}-\mathcal{B}_{\sigma}(\mathbf{D}_{t})$	Render base with restored high-frequency detail

Table 2: RIVR candidates for no-reference in-video restoration.

Method	LPIPS $\downarrow$							CLIP-IQA $\uparrow$							MUSIQ $\uparrow$
	GS2E			GSOTM				GS2E			GSOTM				GS2E			GSOTM
	Slight	Medium	Severe	MB	RS	MB+RS	PN	Slight	Medium	Severe	MB	RS	MB+RS	PN	Slight	Medium	Severe	MB	RS	MB+RS	PN
All-view $G_{0}$	0.397	0.414	0.430	0.323	0.334	0.419	0.476	0.266	0.261	0.256	0.194	0.212	0.196	0.227	34.82	32.02	30.11	36.75	50.25	34.51	34.06
Select-G $G_{0}$ (Ours)	0.386	0.401	0.415	0.285	0.333	0.393	0.472	0.266	0.256	0.261	0.201	0.211	0.187	0.224	36.53	34.51	34.22	42.31	50.55	39.46	35.42
DiFiX3D ( Wu et al. 2025a )	0.349	0.372	0.388	0.224	0.224	0.322	0.249	0.258	0.255	0.253	0.231	0.253	0.213	0.230	41.53	39.03	36.93	41.64	59.77	39.94	51.64
GeoQuery ( Cao et al. 2026 )	0.350	0.373	0.389	0.231	0.224	0.332	0.249	0.256	0.253	0.252	0.230	0.250	0.216	0.231	41.33	38.70	37.20	41.86	59.50	40.02	51.27

Table 3: Perceptual comparison on GS2E and GSOTM at observed input poses. GS2E averages ten scenes per blur level, and GSOTM averages four scenes per degradation type. LPIPS uses sharp targets; CLIP-IQA and MUSIQ are no-reference metrics. Best and second-best results are highlighted.

Variant	PSNR $\uparrow$							SSIM $\uparrow$							LPIPS $\downarrow$
	GS2E			GSOTM				GS2E			GSOTM				GS2E			GSOTM
	S	M	V	MB	RS	MB+RS	PN	S	M	V	MB	RS	MB+RS	PN	S	M	V	MB	RS	MB+RS	PN
RVA w/o cov.	20.123	19.538	19.499	21.621	18.010	17.624	18.413	0.607	0.592	0.589	0.730	0.530	0.534	0.573	0.388	0.413	0.415	0.290	0.339	0.387	0.472
Full RVA	19.934	20.076	20.011	22.545	18.124	18.010	18.458	0.606	0.609	0.608	0.739	0.530	0.541	0.573	0.386	0.401	0.415	0.285	0.333	0.393	0.472

Table 4: Ablation of RVA coverage allocation. S/M/V are GS2E blur levels; MB/RS/MB+RS/PN are GSOTM degradation settings.

Figure 3: Qualitative comparison of renderings at observed input camera poses produced by all-view 3DGS and the Select-view initialization of ClearGS. Red boxes highlight regions where our reconstruction better preserves thin structures, object boundaries, and local appearance relative to all-view supervision.

Figure 4: Qualitative comparison of repair outputs on two GS2E scenes. Each row shows the same camera pose. From left to right, the panels show the common pre-repair 3DGS rendering, the repairs produced by DiFiX3D and GeoQuery, and the repair selected by ClearGS.

Metric	Variant	GS2E			GSOTM
Metric	Variant	S	M	V	MB	RS	MB+RS	PN
LPIPS $\downarrow$	w/o Obs.	0.385	0.401	0.414	0.247	0.294	0.358	0.398
	w/o $F_{t}$	0.336	0.348	0.365	0.171	0.213	0.298	0.196
	Full RIVR	0.333	0.342	0.350	0.193	0.225	0.302	0.197
CLIP-IQA $\uparrow$	w/o Obs.	0.283	0.270	0.279	0.217	0.239	0.209	0.258
	w/o $F_{t}$	0.280	0.275	0.282	0.263	0.280	0.221	0.269

Table 5: Ablation of RIVR candidate selection. “w/o Obs.” replaces the raw video observation

O_{t}

with the current render

R_{t}

as the input to Restormer, while “w/o

F_{t}

” removes the high-frequency fused candidate. The remaining repair-distillation pipeline is unchanged.

Figure 5: Qualitative comparison on five GS2E scenes. From left to right: the input 3DGS, DiFiX3D, GeoQuery, SyncFix, ClearGS, and the ground truth. ClearGS recovers sharper object boundaries and local details while preserving the scene appearance.

3D Gaussian Splatting

3D Gaussian Splatting represents scenes as optimizable 3D Gaussian primitives for fast, photorealistic rendering from new viewpoints.

See all 1,019 papers

CleanGS: Pareto-Guided Post-Hoc Cleanup of 3D Gaussian Splatting Scenes
Jan 1, 2026Jyotirmaya Shivottam, Om Kumar, Subhankar Mishra3D Gaussian Splatting 3D Gaussian Pruning
School of Computer Sciences National Institute of Science Education and Research Bhubaneswar, India
JSGS: JPEG State-Guided Supervision for 3D Gaussian Splatting from Mixed-Quality Views
Aug 9, 2026Jinhua Cui, Anhong Wang, Kai Hu +53D Gaussian Splatting 3D Reconstruction
1Taiyuan University of Science and Technology · 2Chn Energy Digital Intelligence Technology Development (Beijing)
GaussVid: Sparse-View Gaussian Splatting with 3D-Aware Video Diffusion Priors
Aug 22, 2026Xinhui Liu, Can Wang, Wei Jiang +2Sparse-View 3D Reconstruction 3D Gaussian Splatting
School of Computing and Data Science, The University of Hong Kong, Hong Kong, China · Futurewei Technologies Inc, Santa Clara, CA, USA

Jan 1, 2026J/K move · Enter open · S save

CleanGS: Pareto-Guided Post-Hoc Cleanup of 3D Gaussian Splatting Scenes

Jyotirmaya Shivottam, Om Kumar, Subhankar Mishra

School of Computer Sciences National Institute of Science Education and Research Bhubaneswar, India

3D Gaussian Splatting 3D Gaussian Pruning

Open paper AskPDF arXiv

A pretrained 3D Gaussian Splatting (3DGS) reconstruction can reproduce its capture views well while retaining off-object structure that becomes visible under reframing. We present CleanGS, a post-hoc cleanup method that removes such structure without retraining the underlying representation. Given calibrated views and binary target masks, CleanGS scores candidate three-view subsets using mask reliability, camera baseline, viewing-direction diversity, and complementary sparse 3D support. Rather than collapsing these objectives a priori, it computes the exact Pareto front over subsets of the reduced candidate pool and selects a balanced on-dominated subset using weighted Chebyshev scalarization. The selected views then drive rasterization-aware two-of-three voting to retain a subset of the original Gaussians. We evaluate the exported asset directly and introduce a fixed-pose widened-field protocol that exposes structure outside the original framing. Across 18 scenes, CleanGS achieves 19.03 dB standard and 17.83 dB widened-field PSNR, compared with 16.44/16.29 dB for FlashSplat and 15.28/15.10 dB for GaussianCut with the same three masks. CleanGS leads five of six photometric metrics and remains within 0.32 dB of the uncleaned reconstruction in standard PSNR. Its filtering-and-export stage takes a median of 20 seconds and 1.9 GB peak memory, compared with 3.1 minutes and 41.6 GB for FlashSplat and 4.2 minutes and 34.9 GB for GaussianCut. Code is available at https://github.com/smlab-niser/clean-gs/.

Abstract

3D Gaussian Splatting

Explore similar work

CleanGS: Pareto-Guided Post-Hoc Cleanup of 3D Gaussian Splatting Scenes

JSGS: JPEG State-Guided Supervision for 3D Gaussian Splatting from Mixed-Quality Views

GaussVid: Sparse-View Gaussian Splatting with 3D-Aware Video Diffusion Priors