Datasets › PIE-Bench

PIE-Bench (Prompt-based Image Editing Benchmark)

Introduced by Xuan Ju et al. in Direct Inversion: Boosting Diffusion-based Editing with 3 Lines of Code2 Oct 2023 archive 2025-07-28

PIE-Bench comprises 700 images featuring 10 distinct editing types. Images are evenly distributed in natural and artificial scenes (e.g., paintings) among four categories: animal, human, indoor, and outdoor. Each image in PIE-Bench includes five annotations: source image prompt, target image prompt, editing instruction, main editing body, and the editing mask. Notably, the editing mask annotation (indicating the anticipated editing region) is crucial in accurate metrics computations as we expect the editing to only occur within a designated area.

Benchmarks archive 2025-07-28

All 1 leaderboard whose dataset resolves to this page shown (sort by any header). "First row" is the archive's own first row at snapshot, in the archive's row order; nothing here re-ranks and metric direction is not asserted.

First row (archive order)PaperCode
Text-based Image Editing PIE-Bench KV-Edit Background PSNR 35.87 KV-Edit: Training-Free Image Editing for Precise... Xilluill/KV-Edit 18 Compare

Papers archive 2025-07-28

12 shown of 12 papers with a leaderboard row on this dataset's benchmarks, newest first. The archive's own "papers using this dataset" list was never published, so this is the benchmark-backed subset; the archive's count for this dataset is 23. The Syntology column is from Syntology's graph (read 2026-09-24), stated per sample; it is not part of any archive number.

DateSamples run Syntology
KV-Edit: Training-Free Image Editing for Precise Background Preservation 1 1 24 Feb 2025 ran 15 of 31 samples (16 unverified)
FireFlow: Fast Inversion of Rectified Flow for Image Semantic Editing 1 2 10 Dec 2024 ran 4 of 11 samples (7 unverified)
Vision-guided and Mask-enhanced Adaptive Denoising for Prompt-based Image Editing 1 1 14 Oct 2024 not harvested
Inversion-Free Image Editing with Natural Language 1 3 7 Dec 2023 ran 1 of 1 samples (0 unverified; 1 pointer-only for licence)
Direct Inversion: Boosting Diffusion-based Editing with 3 Lines of Code 3 4 2 Oct 2023 ran 4 of 6 samples (2 unverified; 6 pointer-only for licence)
Negative-prompt Inversion: Fast Image Inversion for Editing with Text-guided Diffusion Models 0 1 26 May 2023 not harvested
MasaCtrl: Tuning-Free Mutual Self-Attention Control for Consistent Image Synthesis and Editing 4 1 17 Apr 2023 ran 4 of 7 samples (3 unverified; 1 pointer-only for licence)
StyleDiffusion: Prompt-Embedding Inversion for Text-Based Editing 1 1 28 Mar 2023 ran 1 of 1 samples (0 unverified; 1 pointer-only for licence)
Zero-shot Image-to-Image Translation 2 1 6 Feb 2023 ran 0 of 3 samples (3 unverified)
Plug-and-Play Diffusion Features for Text-Driven Image-to-Image Translation 4 1 22 Nov 2022 ran 1 of 1 samples (0 unverified)
Null-text Inversion for Editing Real Images using Guided Diffusion Models 4 1 17 Nov 2022 ran 1 of 1 samples (0 unverified)
Prompt-to-Prompt Image Editing with Cross Attention Control 7 1 2 Aug 2022 ran 6 of 15 samples (9 unverified; 3 pointer-only for licence)

Dataset loaders archive 2025-07-28

No loader listed in the archive.

Tasks archive 2025-07-28

License archive 2025-07-28

https://docs.google.com/document/d/1K7pjAgGwS8-lsA-D6Me4jAKmYszKpcyydiAEyZuTm2I/edit?usp=sharing

Modalities archive 2025-07-28

Languages archive 2025-07-28

Variants archive 2025-07-28

  • PIE-Bench

1 variant name, as the archive lists them.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections