多奖励作为条件的基于指令的图像编辑
Multi-Reward as Condition for Instruction-based Image Editing
AI总结：
针对基于指令的图像编辑训练数据质量差的问题，提出基于GPT-4o的多视角奖励体系构建RewardEdit20K数据集，设计多奖励条件训练框架，并构建Real-Edit基准，实验证明模型性能优于无奖励版本。
摘要：
高质量的训练三元组（指令、原始图像、编辑后图像）对基于指令的图像编辑至关重要。主流训练数据集（如InsPix2Pix）是使用文本到图像生成模型（如Stable Diffusion、DALL-E）构建的，而这些模型并未针对图像编辑进行训练。因此，这些数据集存在指令遵循不准确、细节保留效果差以及生成伪影的问题。本文提出通过多视角奖励数据而非优化真值图像质量来解决训练数据质量问题。1）首先设计了基于顶尖LVLM（大视觉语言模型，本文采用GPT-4o）的量化指标体系，从指令遵循、细节保留、生成质量3个视角评估生成质量。针对每个视角收集0~5分的量化分数以及关于真值编辑图像中具体失败点的文本描述反馈，构建了高质量编辑奖励数据集RewardEdit20K。2）进一步提出一种新型训练框架，将作为多奖励的指标输出无缝集成到编辑模型中，使模型能从不完美的训练三元组中学习。训练期间，奖励分数和文本描述被编码为嵌入，作为辅助条件输入编辑模型的潜空间和U-Net。3）还构建了包含真实世界图像/照片和多样编辑指令的高难度评估基准Real-Edit。实验表明，在InsPix2Pix和SmartEdit两种主流编辑流水线中，本文的多奖励条件模型均优于无奖励的同类模型。代码已发布于https://github.com/bytedance/Multi-Reward-Editing。
英文摘要：
High-quality training triplets (instruction, original image, edited image) are essential for instruction-based image editing. Predominant training datasets (e.g., InsPix2Pix) are created using text-to-image generative models (e.g., Stable Diffusion, DALL-E) which are not trained for image editing. Accordingly, these datasets suffer from inaccurate instruction following, poor detail preserving, and generation artifacts. In this paper, we propose to address the training data quality issue with multi-perspective reward data instead of refining the ground-truth image quality. 1) we first design a quantitative metric system based on best-in-class LVLM (Large Vision Language Model), i.e., GPT-4o in our case, to evaluate the generation quality from 3 perspectives, namely, instruction following, detail preserving, and generation quality. For each perspective, we collected quantitative score in $0\sim 5$ and text descriptive feedback on the specific failure points in ground-truth edited images, resulting in a high-quality editing reward dataset, i.e., RewardEdit20K. 2) We further proposed a novel training framework to seamlessly integrate the metric output, regarded as multi-reward, into editing models to learn from the imperfect training triplets. During training, the reward scores and text descriptions are encoded as embeddings and fed into both the latent space and the U-Net of the editing models as auxiliary conditions. 3) We also build a challenging evaluation benchmark with real-world images/photos and diverse editing instructions, named Real-Edit. Experiments indicate that our multi-reward conditioned model outperforms its no-reward counterpart on two popular editing pipelines, i.e., InsPix2Pix and SmartEdit. Code is released at https://github.com/bytedance/Multi-Reward-Editing.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
