What question did this study set out to answer?

The study aims to introduce SSD-Edit, a novel framework for instruction-based editing of images and videos using Mamba-2 diffusion.

February 21, 2026Open Access

SSD-Edit: Instruction-Based Image and Video Editing with Mamba-2 Diffusion Backbones

Key Points

The study aims to introduce SSD-Edit, a novel framework for instruction-based editing of images and videos using Mamba-2 diffusion.
Developed an instruction-based image and video editing framework called SSD-Edit.
Introduced Source-Instruction Dual-Stream SSD for handling edits.
Employed Selective Edit Masking via State Divergence for automating edit localization.
Implemented Temporal Edit Propagation for maintaining consistency in video edits.
SSD-Edit effectively integrates instruction-based editing for images and videos.
Demonstrated improved localization of edits compared to previous methods.
Achieved consistent video editing outcomes through temporal propagation of edits.

Abstract

We propose SSD-Edit, the first instruction-based image and video editing framework built on Mamba-2 diffusion. No prior work has applied any Mamba variant to instruction-following editing. SSD-Edit introduces Source-Instruction Dual-Stream SSD, Selective Edit Masking via State Divergence for automatic edit localization, and Temporal Edit Propagation for consistent video editing.

SSD-Edit: Instruction-Based Image and Video Editing with Mamba-2 Diffusion Backbones

Key Points

Abstract

Cite This Study