AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

A Loss Function for Generative Neural Networks Based on Watson’s Perceptual Model

Steffen Czolbe, Oswin Krause, Ingemar Cox, Christian Igel

NeurIPS 2020 · Image generation › Autoregressive and latent generation · approach: Classical generative image synthesis · 16 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use Watson-DFT instead of LPIPS-VGG as the lossColor-image loss computation · Typical learning scenarioRuntime: $56.8$ s vs. $16.2$ sImproved
Use Watson-DFT instead of LPIPS-VGG as the lossColor-image loss computation · Typical learning scenarioGPU memory: $2214$ Mb vs. $114$ MbImproved
Train the VAE with Watson-DFT rather than L2 or SSIMMNIST-VAE · MNISTVisual sample qualityImproved
Train the VAE with LPIPS-VGG or LPIPS-Squeeze rather than Watson-DFTMNIST-VAE · MNISTVisual sample qualityHurt
Train the VAE with Watson-DFT rather than SSIM or Adaptive-LosscelebA-VAE · celebAVisual sample qualityImproved
Train the VAE with LPIPS-VGG rather than Watson-DFTcelebA-VAE · celebAFine-detail generationMixed
Train the VAE with LPIPS-Squeeze rather than Watson-DFTcelebA-VAE · celebAVisual artifactsHurt
Use Watson-DFT rather than L2, SSIM, or Adaptive-LossPerceptual similarity metric · 2AFC validationAgreement with human judgmentsImproved
Use LPIPS-VGG or LPIPS-Squeeze rather than Watson-DFTPerceptual similarity metric · 2AFC validationAgreement with human judgmentsImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy