ImVideoEdit: Image-learning Video Editing via 2D Spatial Difference Attention Blocks
This addresses scalability issues in video editing for practitioners by reducing data and computational requirements, though it is incremental as it builds on existing spatiotemporal decoupling ideas.
The paper tackled the problem of expensive paired video data for video editing by proposing ImVideoEdit, a framework that learns editing from image pairs, achieving comparable fidelity and temporal consistency to larger models trained on extensive video datasets with only 13K image pairs and low computational overhead.
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.