DiffSeg30k: A Multi-Turn Diffusion Editing Benchmark for Localized AIGC Detection

ICLR 2026 Conference Submission7577 Authors

Keywords: AIGC Detection, Diffusion Editing, Segmentation

TL;DR: A dataset to support fine-grained segmentation of diffusion-edited local areas.

Abstract:

Diffusion-based editing enables realistic modification of local image regions, making AI-generated content harder to detect. Existing AIGC detection benchmarks focus on classifying entire images, overlooking the localization of diffusion-based edits.
We introduce DiffSeg30k, a publicly available dataset of 30k diffusion-edited images with pixel-level annotations, designed to support fine-grained detection. DiffSeg30k features: 1) In-the-wild images—we collect images or image prompts from COCO to reflect real-world content diversity; 2) Diverse diffusion models—local edits using eight SOTA diffusion models; 3) Multi-turn editing—each image undergoes up to three sequential edits to mimic real-world sequential editing; and 4) Realistic editing scenarios—a vision-language model (VLM)-based pipeline automatically identifies meaningful regions and generates context-aware prompts covering additions, removals, and attribute changes.
DiffSeg30k shifts AIGC detection from binary classification to semantic segmentation, enabling simultaneous localization of edits and identification of the editing models. We benchmark two baseline segmentation approaches, revealing significant challenges in segmentation tasks, particularly concerning robustness to image distortions.
We believe DiffSeg30k will advance research in fine-grained localization of AI-generated content by demonstrating the promise and limitations of segmentation-based detection methods.

Primary Area: datasets and benchmarks

Submission Number: 7577