Video object removal and injection are reciprocal editing operations between object-present and object-absent video states, yet they differ substantially in training difficulty. Removal can be effectively bootstrapped from video inpainting and further adapted to erase object-related interaction effects. Injection, however, requires large-scale paired supervision to model object appearance, dynamics, and scene-dependent effects, while perfectly aligned real-world pairs are difficult to collect and synthetic data lacks sufficient realism and diversity. To address this challenge, we present UniVOE, a removal-driven framework for training both video object removal and injection models. We first train an effect-aware remover by transferring video inpainting priors and finetuning on object-present/object-absent pairs. The trained remover converts real object-present videos into object-absent counterparts, creating large-scale paired supervision for injection. An injection model is trained on these pairs in the reverse direction to recover objects together with their dynamics and interaction effects. Moreover, since pixel-level reconstruction alone cannot adequately capture perceptual realism, we introduce a reward-based post-training method that aligns injection results with human preferences using a proprietary reward model. Experiments against strong open-source and commercial systems demonstrate that UniVOE achieves state-of-the-art performance on both object removal and injection, particularly in challenging interaction-effect cases.
Figures & tables
Figure 2: UniVOE uses removal to scale paired video data up, drives the training of injection.
Figure 3: Workflow of UniVOE. (a) The remover and injector share the same network, but use different instructions and VOE units. (b) For removal, VOE unit contains an input video and an inflated binary mask video to indicate object to be removed. For injection, an all-one mask is used to enable global interaction-aware injection. The reference image is optional in removal and injection. It is generated with external image editing model, providing more specific editing guidance. (c) To improve physical realism of injected results, we propose GRAR to update the pre-trained injection model toward human preferences.
Figure 4: Visual comparison on video object removal. Other methods either remain object-related effects traces ( e.g. , lighting of the car), or predict unnatural background contents. In contrast, our method removes object itself and recovers physically realistic scene. It outperforms both open-source and commercial models. More visual comparison can be found in supplementary.
Data
Metric
VidPaint
VACE
Minimax
EffErase
Bernini
Kling3
Seedance2
Ours
VOR
PSNR ↑
17.425
11.567
17.243
23.210
16.380
21.581
20.002
23.752
SSIM ↑
0.705
0.601
0.692
0.734
0.674
0.712
0.715
0.786
LPIPS ↓
0.253
0.441
0.229
0.142
0.247
0.182
0.204
0.130
FVD ↓
563.08
825.69
171.66
150.82
238.47
210.58
174.36
146.37
ROSE
PSNR ↑
25.693
15.377
28.839
24.666
23.814
22.433
27.286
29.596
SSIM ↑
0.896
0.388
0.745
0.918
0.876
0.903
0.912
0.919
Table 1: Comparison on removal. The best results are highlighted in bold . Video materials are given in supplementary.
Data
Metric
VidPaint
VACE
UniVideo
Bernini
Kling3
Seedance2
Ours
VOR
DINO-I ↑
0.310
0.232
0.387
0.327
0.303
0.381
0.390
CLIP-P ↑
0.224
0.200
0.228
0.208
0.227
0.217
0.232
LPIPS ↓
0.257
0.366
0.314
0.329
0.295
0.249
0.237
FVD ↓
391.63
505.02
300.69
337.48
296.67
259.31
234.56
ROSE
DINO-I ↑
0.241
0.192
0.346
0.319
0.312
0.334
0.347
CLIP-P ↑
0.220
0.217
0.236
0.242
0.221
0.243
0.248
Table 2: Comparison on injection. The best results are highlighted in bold . Video materials are given in supplementary.
Figure 5: Visual comparison on video object injection. Corresponding prompt is given above. Other methods either inject background-inconsistent object, or fail to inject harmonious object. Our model predicts matching target and credible dynamics.
Figure 6: Visual ablation of training data used for the remover. “w/o inpainting” is trained by noisy removal pairs and carries the risk of overall color shifts. “UniVOE-” introduces inpainting data to avoid excessive background editing.
Settings
Inpaint
Removal
PSNR ↑
SSIM ↑
LPIPS ↓
FVD ↓
w/o inpainting
0K
68K
20.578
0.711
0.158
163.130
UniVOE-
34K
34K
22.891
0.740
0.141
149.777
Table 3: Ablation on inpainting data for training remover.
Settings
DINO-I ↑
CLIP-P ↑
LPIPS ↓
FVD ↓
w/o GRAR
0.353
0.221
0.307
361.286
UniVOE
0.390
0.232
0.237
234.562
Table 4: Ablation on post-training stage for object injection.
Appendix figures & tables5 assets
Supplementary material from the paper’s appendix.
Appendix
Settings
DINO-I ↑
CLIP-P ↑
LPIPS ↓
FVD ↓
w/ DiffusionNFT
0.388
0.229
0.289
314.101
w/ GRAR
0.387
0.231
0.287
283.812
Appendix
Table 5: Post-training ablation for video object injection. Evaluate on VOR.
Figure 7: Visual comparison of models trained by GRAR and DiffusionNFT. One can see that their outputs are visually consistent, which is consistent with the quantitative results in Tab. 5 . This further demonstrates that the performance of the proposed GRAR approximates that of DiffusionNFT.
Figure 8: Reward curves of GRAR and DiffusionNFT on the smaller injectors with Wan2.1-1.3B backbone.
Figure 9: Ablation on the post-training stage of injector. We showcase 4 cases from input video and the edited results of “w/o RL” and “UniVOE”, respectively. Related prompts are given below. Results of “w/o RL” exhibit color mutation (top-left), subject disappear (top-right), subject distortion (bottom-left), and semantic inconsistency (bottom-right). These issues can be well addressed in common scenarios by post-training.
Figure 10: More results of video object removal and injection of UniVOE. For removal, UniVOE can fill realistic background content for the removed region. For injection, UniVOE not only inserts desired objects, but also predicts its natural interaction effects with background. Source videos can be found in supplementary materials.