AI Paper Atlas

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

Object-Conditioned Energy-Based Attention Map Alignment in Text-to-Image Diffusion Models

Yasi Zhang, Peiyu Yu, Ying Nian Wu

ECCV 2024 · Image generation › Text-to-image diffusion · approach: Compositional text-to-image guidance · 13 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 the repulsive (negative sampling) term to the binding loss (ablation)EBAMA lambda=0 · AnE dataset Object-ObjectT-C Sim.: 0.765 -> 0.823Improved
add the repulsive term to the full method (ablation)EBAMA · AnE dataset Object-ObjectT-C Sim.: 0.819 -> 0.836Improved
condition the repulsive energy on objects rather than averaging object and modifier energy (ablation)EBAMA lambda=0 · AnE dataset Animal-ObjectT-C Sim.: 0.832 -> 0.851Improved
object conditioning on Animal-Animal prompts with no modifiers (ablation)EBAMA lambda=0 · AnE dataset Animal-AnimalT-C Sim.: 0.814 -> 0.814No clear effect
use cosine similarity energy instead of average KL divergence (ablation)EBAMA lambda=0 · AnE dataset Object-ObjectT-C Sim.: 0.817 -> 0.823Improved
add object-centric intensity regularizer (lambda>0) on harder prompts (ablation)EBAMA · AnE dataset Object-ObjectT-C Sim.: 0.823 -> 0.836Improved
set intensity weight lambda to 1.0 (ablation)EBAMA · qualitative promptsimage quality / bindingHurt
update latents at all 50 steps instead of the first 25 (ablation)EBAMA · qualitative promptsartifactsHurt
append more unrelated external modifiers (n=1 to 3) (ablation)EBAMA · AnE dataset Animal-AnimalFull Sim.: 0.340 -> 0.331Hurt
replace SG-style pair loss with EBAMA binding lossStable Diffusion v1.4 · AnE dataset Animal-AnimalT-C Sim.: 0.767 -> 0.817Improved
use EBAMA instead of AnE for inference-time latent updateStable Diffusion v1.4 · ABC-6K 100 promptsgeneration time (min): 47.0 -> 20.9Improved

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