AI Paper Atlas

Research map / Image generation / Editing and style transfer

Cross Attention Based Style Distribution for Controllable Person Image Synthesis

Xinyue Zhou, Mingyu Yin, Xinyuan Chen, Li Sun, Changxin Gao, Qingli Li

ECCV 2022 · Image generation › Editing and style transfer · approach: Pose-guided person image synthesis · 75 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 self-attention on style features in CASD block (ablation)CASD · DeepFashion pose transferFID: 13.2462 -> 11.3732Improved
add self-attention on style features in CASD block (ablation)CASD · DeepFashion pose transferSSIM: 0.7201 -> 0.7248Improved
add attention matrix prediction AM_p from target pose (ablation)CASD · DeepFashion pose transferFID: 12.4265 -> 11.3732Improved
add attention matrix prediction AM_p from target pose (ablation)CASD · DeepFashion pose transferLPIPS: 0.1985 -> 0.1936Improved
add AMCE loss guiding cross attention with parsing map (ablation)CASD · DeepFashion pose transferFID: 14.7513 -> 11.3732Improved
add AMCE loss guiding cross attention with parsing map (ablation)CASD · DeepFashion pose transferSSIM: 0.7156 -> 0.7248Improved
add AMCE loss guiding cross attention with parsing map (ablation)CASD · DeepFashion pose transferPSNR: 31.41 -> 31.67Improved
use CASD instead of GFLApose transfer · DeepFashion pose transferFID: 10.5730 -> 11.3732Hurt
use CASD instead of SPGNetpose transfer · DeepFashion pose transferuser study Jab: 17.26% -> 28.96%Improved
use CASD instead of PISE for virtual try-on (upper cloth)try-on without retraining · Virtual try-on upper clothFID: 14.0537 -> 12.5376Improved
use CASD instead of SPGNet for parsing predictionparsing map · DeepFashion parsingaverage IoU: 61.89 -> 66.34Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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