AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Learning Prior Feature and Attention Enhanced Image Inpainting

Chenjie Cao, Qiaole Dong, Yanwei Fu

ECCV 2022 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Masked autoencoder pretraining · 47 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 MAE prior features with naive element-wise addition (ablation)LaMa-style ACR · Places2 5-scene subset 256x256FID: 28.01 -> 28.01No clear effect
multiply MAE prior features by a zero-initialized learnable alpha before addition (ablation)ACR + MAE · Places2 5-scene subset 256x256FID: 28.01 -> 26.83Improved
initialize alpha with 1 instead of 0 (ablation)ACR + MAE · Places2 5-scene subset 256x256FID: 26.83 -> 26.90No clear effect
add trainable contextual attention (ablation)ACR + MAE · Places2 5-scene subset 256x256FID: 26.84 -> 26.99Hurt
add prior attention aggregation from MAE decoder (ablation)ACR + MAE (mixed masks) · Places2 5-scene subset 256x256FID: 26.84 -> 25.49Improved
pre-train MAE with mixed continuous masks instead of random masks (ablation)FAR with prior attention · Places2 5-scene subset 256x256FID: 26.25 -> 25.49Improved
finetune MAE for partially masked patches (ablation)FAR full · Places2 5-scene subset 256x256FID: 25.49 -> 26.00Hurt
add 2D coordinate grid to resized prior features (ablation)FAR · Places2 512x512 finetunedFID: 25.47 -> 25.33Improved
remove patch-wise pixel normalization from MAE target (ablation)MAE pre-training · Places2 512x512quality: 25.33 -> 25.30Improved
replace LaMa with FAR on Places2 256LaMa · Places2 256x256FID: 1.63 -> 1.31Improved
replace LaMa with FAR on Places2 512LaMa · Places2 512x512FID: 27.8 -> 25.3Improved
replace LaMa with FAR on FFHQ 256LaMa · FFHQ 256x256FID: 6.38 -> 6.15Improved

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