UltraEdit: Instruction-based Fine-Grained Image Editing at Scale
This addresses the need for diverse, high-quality training data for image editing models, though it is incremental as it builds on existing datasets like InstructPix2Pix and MagicBrush.
The paper tackles the problem of limited and biased datasets for instruction-based image editing by introducing UltraEdit, a large-scale dataset of approximately 4 million automatically generated editing samples, which sets new records on MagicBrush and Emu-Edit benchmarks.
This paper presents UltraEdit, a large-scale (approximately 4 million editing samples), automatically generated dataset for instruction-based image editing. Our key idea is to address the drawbacks in existing image editing datasets like InstructPix2Pix and MagicBrush, and provide a systematic approach to producing massive and high-quality image editing samples. UltraEdit offers several distinct advantages: 1) It features a broader range of editing instructions by leveraging the creativity of large language models (LLMs) alongside in-context editing examples from human raters; 2) Its data sources are based on real images, including photographs and artworks, which provide greater diversity and reduced bias compared to datasets solely generated by text-to-image models; 3) It also supports region-based editing, enhanced by high-quality, automatically produced region annotations. Our experiments show that canonical diffusion-based editing baselines trained on UltraEdit set new records on MagicBrush and Emu-Edit benchmarks. Our analysis further confirms the crucial role of real image anchors and region-based editing data. The dataset, code, and models can be found in https://ultra-editing.github.io.