The reviewed record of science sign in
Pith

arxiv: 2104.11222 · v3 · pith:PSGSJ7FF · submitted 2021-04-22 · cs.CV · cs.GR· cs.LG

On Aliased Resizing and Surprising Subtleties in GAN Evaluation

Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel pith:PSGSJ7FFrecord.jsonopen to challenge →

classification cs.CV cs.GRcs.LG
keywords imagecompressiongenerativeimagesprocessingresizingvariationsaliasing
0
0 comments X
read the original abstract

Metrics for evaluating generative models aim to measure the discrepancy between real and generated images. The often-used Frechet Inception Distance (FID) metric, for example, extracts "high-level" features using a deep network from the two sets. However, we find that the differences in "low-level" preprocessing, specifically image resizing and compression, can induce large variations and have unforeseen consequences. For instance, when resizing an image, e.g., with a bilinear or bicubic kernel, signal processing principles mandate adjusting prefilter width depending on the downsampling factor, to antialias to the appropriate bandwidth. However, commonly-used implementations use a fixed-width prefilter, resulting in aliasing artifacts. Such aliasing leads to corruptions in the feature extraction downstream. Next, lossy compression, such as JPEG, is commonly used to reduce the file size of an image. Although designed to minimally degrade the perceptual quality of an image, the operation also produces variations downstream. Furthermore, we show that if compression is used on real training images, FID can actually improve if the generated images are also subsequently compressed. This paper shows that choices in low-level image processing have been an underappreciated aspect of generative modeling. We identify and characterize variations in generative modeling development pipelines, provide recommendations based on signal processing principles, and release a reference implementation to facilitate future comparisons.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. iTryOn: Mastering Interactive Video Virtual Try-On with Spatial-Semantic Guidance

    cs.CV 2026-05 unverdicted novelty 7.0

    iTryOn is a video diffusion Transformer that injects spatial 3D hand guidance and semantic action captions to enable interactive garment replacement in videos.

  2. iTryOn: Mastering Interactive Video Virtual Try-On with Spatial-Semantic Guidance

    cs.CV 2026-05 unverdicted novelty 7.0

    iTryOn is a diffusion-based framework that adds spatial 3D hand guidance and semantic action-aware embeddings to handle complex garment deformations during human-clothing interactions in videos.

  3. High-Resolution Image Synthesis with Latent Diffusion Models

    cs.CV 2021-12 conditional novelty 7.0

    Latent diffusion models achieve state-of-the-art inpainting and competitive results on unconditional generation, scene synthesis, and super-resolution by performing the diffusion process in the latent space of pretrai...

  4. Diffusion Models Beat GANs on Image Synthesis

    cs.LG 2021-05 accept novelty 7.0

    Diffusion models with architecture improvements and classifier guidance achieve superior FID scores to GANs on unconditional and conditional ImageNet image synthesis.

  5. Latte: Latent Diffusion Transformer for Video Generation

    cs.CV 2024-01 unverdicted novelty 6.0

    Latte achieves state-of-the-art video generation on FaceForensics, SkyTimelapse, UCF101, and Taichi-HD by using a latent diffusion transformer with four efficient spatial-temporal decomposition variants and best-pract...