AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

Label-free Neural Semantic Image Synthesis

Jiayi Wang, Kevin A Laube, Yumeng Li, Jan Hendrik Metzen, Shin-I Cheng, Julio Borges, Anna Khoreva

ECCV 2024 · Image generation › Text-to-image diffusion · approach: Compositional text-to-image guidance · 2 citations (OpenAlex)

Read the paper

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 neural layout (20 PCA) instead of ground-truth semantic segmentation as ControlNet conditioningControlNet · COCO-StuffmIoU: 43.3 -> 52.9Improved
use SD features rather than CLIP for neural layout (ablation)LUMEN · COCO-StuffmIoU: 41.7 -> 51.4Improved
increase number of PCA components from 1 to 100 (ablation)LUMEN (SD features) · COCO-StuffmIoU: 45.0 -> 54.6Improved
increase number of PCA components from 1 to 100 (ablation)LUMEN (SD features) · COCO-StuffLPIPS diversity: 0.53 -> 0.25Hurt
increase number of PCA components from 1 to 100 (ablation)LUMEN (SD features) · COCO-StuffTIFA: 0.87 -> 0.47Hurt
add unlabeled Cityscapes images (L + U) when training the generator with neural layout (ablation)LUMEN · CityscapesmIoU: 57.6 -> 58.2Improved
add unlabeled images using pseudo-labels for Sem. Seg. conditioning (ablation)ControlNet Sem. Seg. · CityscapesmIoU: 57.9 -> 53.3Hurt
add LUMEN synthetic data to real NYUv2 training datamulti-task ResNet-50 · NYUv2 sem. seg.mIoU: 44.6 -> 47.3Improved
add LUMEN synthetic data to real NYUv2 training datamulti-task ResNet-50 · NYUv2 depthAbsErr: 1.17 -> 1.25Hurt
augment Cityscapes training with LUMEN prompt-edited imagesSegFormer · Cityscapes to ACDCAvg. mIoU: 47.0 -> 52.9Improved
use PnP-Diffusion prompt editing for augmentationSegFormer · Cityscapes to ACDCAvg. mIoU: 47.0 -> 48.7Improved
use neural layout instead of Canny edges, depth or predicted Sem. Seg.ControlNet · COCO-StuffTIFA: 0.88 -> 0.66Hurt

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