HQ-Edit: A High-Quality Dataset for Instruction-based Image Editing

TL;DR

HQ-Edit is a large, high-quality dataset for instruction-based image editing created using advanced foundation models, significantly improving model performance and surpassing human-annotated data in quality.

Contribution

The paper presents a scalable pipeline for creating high-quality image editing datasets using GPT-4V and DALL-E 3, and introduces new evaluation metrics for assessing edit quality.

Findings

HQ-Edit dataset contains 200,000 high-quality edits.

Finetuning InstructPix2Pix on HQ-Edit achieves state-of-the-art results.

Proposed metrics effectively evaluate image editing quality.

Abstract

This study introduces HQ-Edit, a high-quality instruction-based image editing dataset with around 200,000 edits. Unlike prior approaches relying on attribute guidance or human feedback on building datasets, we devise a scalable data collection pipeline leveraging advanced foundation models, namely GPT-4V and DALL-E 3. To ensure its high quality, diverse examples are first collected online, expanded, and then used to create high-quality diptychs featuring input and output images with detailed text prompts, followed by precise alignment ensured through post-processing. In addition, we propose two evaluation metrics, Alignment and Coherence, to quantitatively assess the quality of image edit pairs using GPT-4V. HQ-Edits high-resolution images, rich in detail and accompanied by comprehensive editing prompts, substantially enhance the capabilities of existing image editing models. For example, an HQ-Edit finetuned InstructPix2Pix can attain state-of-the-art image editing performance, even surpassing those models fine-tuned with human-annotated data. The project page is https://thefllood.github.io/HQEdit_web.