Edit2Interp: Adapting Image Foundation Models from Spatial Editing to Video Frame Interpolation with Few-Shot Learning
Authors: Nasrin Rahimi, Mısra Yavuz, Burak Can Biner, Yunus Bilge Kurt, Ahmet Rasim Emirdağı, Süleyman Aslan, Görkay Aydemir, M. Akın Yılmaz, A. Murat Tekalp
Organizations: Codeway AI Research · Dept. of Electrical & Electronics Engineering, Koç University, İstanbul, Türkiye
Pre-trained image editing models exhibit strong spatial reasoning and object-aware transformation capabilities acquired from billions of image-text pairs, yet they possess no explicit temporal modeling. This paper demonstrates that these spatial priors can be repurposed to unlock temporal synthesis capabilities through minimal adaptation - without introducing any video-specific architecture or motion estimation modules. We show that a large image editing model (Qwen-Image-Edit), originally designed solely for static instruction-based edits, can be adapted for Video Frame Interpolation (VFI) using only 64-256 training samples via Low-Rank Adaptation (LoRA). Our core contribution is revealing that the model's inherent understanding of "how objects transform" in static scenes contains latent temporal reasoning that can be activated through few-shot fine-tuning. While the baseline model completely fails at producing coherent intermediate frames, our parameter-efficient adaptation successfully unlocks its interpolation capability. Rather than competing with task-specific VFI methods trained from scratch on massive datasets, our work establishes that foundation image editing models possess untapped potential for temporal tasks, offering a data-efficient pathway for video synthesis in resource-constrained scenarios. This bridges the gap between image manipulation and video understanding, suggesting that spatial and temporal reasoning may be more intertwined in foundation models than previously recognized
Figures & tables
Figure 1: Conceptual comparison of VFI paradigms. (a) Traditional VFI networks,(b) Task-specific diffusion models, and (c) Our Edit2Interp approach
Figure 2: Overview of the Edit2Interp framework
Dataset
Method
PSNR ↑
LPIPS ↓
FID ↓
FloLPIPS ↓
PS ↑
VIMEO90k
Baseline (image edit)
19.80
0.207
11.26
0.258
51.53
LoRA-r64-N64
23.63
0.074
4.38
0.104
88.29
LoRA-r64-N128
26.17
0.053
3.50
0.766
100.35
LoRA-r128-N64
24.01
0.066
3.94
0.095
90.59
LoRA-r128-N128
26.66
0.051
3.47
0.072
101.04
Table 1: Quantitative comparison on different datasets.best in bold, second best underlined.
Figure 3: Visual comparison across datasets. Top row for each dataset: ground truth frames t−1 , t , t+1 , baseline interpolation t , and LoRA interpolation t . Bottom row for each dataset: absolute difference maps with respect to the ground truth frame t .
Current video editing models often rely on expensive paired video data, which limits their practical scalability. In essence, most video editing tasks can be formulated as a decoupled spatiotemporal process, where the temporal dynamics of the pretrained model are preserved while spatial content is selectively and precisely modified. Based on this insight, we propose ImVideoEdit, an efficient framework that learns video editing capabilities entirely from image pairs. By freezing the pre-trained 3D attention modules and treating images as single-frame videos, we decouple the 2D spatial learning process to help preserve the original temporal dynamics. The core of our approach is a Predict-Update Spatial Difference Attention module that progressively extracts and injects spatial differences. Rather than relying on rigid external masks, we incorporate a Text-Guided Dynamic Semantic Gating mechanism for adaptive and implicit text-driven modifications. Despite training on only 13K image pairs for 5 epochs with exceptionally low computational overhead, ImVideoEdit achieves editing fidelity and temporal consistency comparable to larger models trained on extensive video datasets.
Large pre-trained video diffusion models excel in video frame interpolation but struggle to generate high fidelity frames due to reliance on intrinsic generative priors, limiting detail preservation from start and end frames. Existing methods often depend on motion control for temporal consistency, yet dense optical flow is error-prone, and sparse points lack structural context. In this paper, we propose FC-VFI for faithful and consistent video frame interpolation, supporting 4×x and 8× interpolation, boosting frame rates from 30 FPS to 120 and 240 FPS at 2560×1440resolution while preserving visual fidelity and motion consistency. We introduce a temporal modeling strategy on the latent sequences to inherit fidelity cues from start and end frames and leverage semantic matching lines for structure-aware motion guidance, improving motion consistency. Furthermore, we propose a temporal difference loss to mitigate temporal inconsistencies. Extensive experiments show FC-VFI achieves high performance and structural integrity across diverse scenarios.
Ganggui Ding, Hao Chen, Xiaogang Xu
⋆Zhejiang University · † Chinese University of Hong Kong
Instruction-based video editing is commonly built on video-pretrained generative backbones: a video diffusion transformer is adapted, at considerable cost, to condition on a source video and an editing instruction. In this report we explore a different route and show that a strong instruction-based image editing model can edit videos by operating directly on video-VAE latents. Starting from Qwen-Image-Edit, we arrange the latent frames of a Wan2.1 video VAE as tiles of one large virtual image, reuse the editor's image positional encoding for every tile, and bridge the two latent spaces with a pair of lightweight input/output projections warm-started from the editor's own patchify and unpatchify layers, so that at initialization a (static) video is embedded exactly as an image the model already understands. The whole system is then fine-tuned on the public Ditto-1M editing triplets, and a few denoising steps of Wan2.2 serve as an optional temporal enhancer. We motivate the design with a chain of zero-training observations: the stock image editor already edits a video presented as a contact sheet; it is indifferent to whether the sheet's tokens come from one joint encode or from per-frame encodes stitched in latent space; and it even edits genuine video latents zero-shot to a clearly recognizable degree, leaving fine-tuning only a fidelity gap to close. Our results suggest that, despite the large investment in training video latent spaces, per-frame video latents remain close enough to the image domain that mature image editing priors transfer with minimal adaptation. Project Page: https://yunpeng1998.github.io/Qwen-Video-Edit-Page Code: https://github.com/yunpeng1998/Qwen-Video-Edit Model: https://huggingface.co/yunpeng1998/Qwen-Video-Edit