AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Evaluation Metrics for Conditional Image Generation

Yaniv Benny, Tomer Galanti, Sagie Benaim, Lior Wolf

IJCV 2021 · Image generation › Autoregressive and latent generation · approach: Synthetic training data from generators · 64 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
Add conditional components (BCIS) instead of unconditional ISSGAN · CIFAR10IS vs BCIS: IS 6.84 vs BCIS 2.71Improved
Use BCIS rather than accuracy to judge conditioningInfoGAN · CIFAR10accuracy vs BCIS: Acc 65.69 vs BCIS 3.25Improved
Measure conditional metrics instead of FID aloneSGAN · MNISTFID vs WCFID: FID 69.34 vs WCFID 609.48Improved
Apply ACGAN auxiliary classifier over CGANCGAN · MNISTWCFID: 98.48 -> 91.21Improved
Apply ACGAN auxiliary classifier over CGANCGAN · CIFAR10BCFID: 6.30 -> 5.25Improved
Longer BigGAN training (100K to full)BigGAN · ImageNetWCFID: 11.36 -> 9.17 -> 6.33Improved
Longer BigGAN training (100K to full)BigGAN · ImageNetBCIS: 12.19 -> 33.71 -> 72.15Improved
Longer BigGAN training (100K to full)BigGAN · ImageNetaccuracy: 18.87 -> 27.70 -> 51.66Improved
Label noising of conditionsIS and FID metrics · MNIST (simulation)IS/FID: unconditional IS and FID constant while BCIS falls and WCIS risesMixed
Mode collapse simulationBCFID vs WCFID · MNIST (simulation)sensitivity: WCFID most noticeable, BCFID less detectableMixed
Use WCFID rather than FID for class-specific mode collapseFID · MNIST (simulation)sensitivity: FID as sensitive as BCFID, less than WCFIDImproved
Image noise, blur, swirl, PCABCIS vs IS · MNIST (simulation)IS: BCIS declines more rapidly than ISImproved

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