mlsanityData to 14 Oct 2026

Rethinking FID Through the Geometry of the Reference Dataset

arXiv 2605.29335 ↗First posted 28 May 2026cs.CV cs.AIcited by 1

Not on the map. Why?

Abstract (arXiv)

Fr'echet Inception Distance (FID) is widely used to evaluate image generators, yet lower FID does not always correspond to better sample quality. We show that this mismatch depends in part on the geometry of the reference dataset. In a controlled study across six datasets, distributional density and effective rank significantly explain how FID changes as sample quality improves. Concentrated datasets tend to yield more favorable FID trends, whereas more dispersed datasets can make FID worsen despite better samples. Attribution to precision and recall and ablations with alternative feature spaces and distances support the same conclusion. These results suggest that distributional metrics should be interpreted together with the geometry of the reference dataset for more reliable benchmarking.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 24 of its 39 references that are in the map. First posted on arXiv 28 May 2026.

Builds on

24 of 39 references.

References · 24

  1. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  2. 2014Microsoft COCO: Common Objects in Contextcited by 17,608
  3. 2020Denoising Diffusion Implicit Modelscited by 9,781
  4. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  5. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,585
  6. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understandingcited by 6,355
  7. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 6,013
  8. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,475
  9. 2017Progressive Growing of GANs for Improved Quality, Stability, and Variationcited by 2,988
  10. 2018Demystifying MMD GANscited by 1,293
  11. 2023ImageReward: Learning and Evaluating Human Preferences for Text-to-Image Generationcited by 1,045
  12. 2019Improved Precision and Recall Metric for Assessing Generative Modelscited by 944
12 more
  1. 2022DPM-Solver++: Fast Solver for Guided Sampling of Diffusion Probabilistic Modelscited by 938
  2. 2021Datasets: A Community Library for Natural Language Processingcited by 530
  3. 2023Rethinking FID: Towards a Better Evaluation Metric for Image Generationcited by 310
  4. 2024Playground v2.5: Three Insights towards Enhancing Aesthetic Quality in Text-to-Image Generationcited by 256
  5. 2022The Role of ImageNet Classes in Fr'echet Inception Distancecited by 213
  6. 2023Exposing flaws of generative model evaluation metrics and their unfair treatment of diffusion modelscited by 205
  7. 2019Effectively Unbiased FID and Inception Score and where to find themcited by 165
  8. 2021The Benchmark Lotterycited by 136
  9. 2023Toward Verifiable and Reproducible Human Evaluation for Text-to-Image Generationcited by 68
  10. 2025Enhanced Diffusion Sampling via Extrapolation with Multiple ODE Solutionscited by 8
  11. 2025MixDiT: Accelerating Image Diffusion Transformer Inference with Mixed-Precision MX Quantizationcited by 8
  12. 2025Tortoise and Hare Guidance: Accelerating Diffusion Model Inference with Multirate Integrationcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026What Does Fr'echet Distance Measure? A Directional Decompositioncited by 0