RefEdit: A Benchmark and Method for Improving Instruction-based Image Editing Model on Referring Expressions
This addresses a key limitation in image editing for AI applications, offering a scalable solution for handling complex scenes, though it is incremental in improving existing methods.
The paper tackles the problem of instruction-based image editing models struggling with complex scenes containing multiple entities by introducing RefEdit, a model trained on synthetic data, which outperforms baselines trained on millions of samples and achieves state-of-the-art results comparable to closed-source methods.
Despite recent advances in inversion and instruction-based image editing, existing approaches primarily excel at editing single, prominent objects but significantly struggle when applied to complex scenes containing multiple entities. To quantify this gap, we first introduce RefEdit-Bench, a rigorous real-world benchmark rooted in RefCOCO, where even baselines trained on millions of samples perform poorly. To overcome this limitation, we introduce RefEdit -- an instruction-based editing model trained on our scalable synthetic data generation pipeline. Our RefEdit, trained on only 20,000 editing triplets, outperforms the Flux/SD3 model-based baselines trained on millions of data. Extensive evaluations across various benchmarks demonstrate that our model not only excels in referring expression tasks but also enhances performance on traditional benchmarks, achieving state-of-the-art results comparable to closed-source methods. We release data \& checkpoint for reproducibility.