ParallelEdits: Efficient Multi-Aspect Text-Driven Image Editing with Attention Grouping

Multiple edits in images pose a significant challenge in existing models (such as DirectInverison [1] and InfEdit [2]), as their performance downgrades with an increasing number of aspects. In contrast, our ParallelEdits can achieve precise multi-aspect image editing in 5 seconds. The symbol denotes a swap action, the symbol denotes an object addition action, and the symbol denotes an object deletion.