AI Paper Atlas

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

Stable Diffusion Models are Secretly Good at Visual In-Context Learning

Trevine Oorloff, Vishwanath Sindagi, Wele Gedara Chaminda Bandara, Ali Shafahi, Amin Ghiasi, Charan Prakash, Reza Ardekani

ICCV 2025 · Image generation › Text-to-image diffusion · approach: Compositional text-to-image guidance · 4 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
use implicitly-weighted prompt ensembling (IWPE) with 5 prompts instead of 1 prompt (ablation)SD-VICL · Pascal-5i foreground segmentationmIoU: 43.92 -> 55.49Improved
use IWPE instead of SegGPT feature ensembling (5 prompts) (ablation)SD-VICL · Cityscapes semantic segmentationmIoU: 22.97 -> 27.01Improved
use IWPE instead of feature ensembling on simpler single-class tasks (ablation)SD-VICL · tasks other than semantic segmentationtask metricsNo clear effect
use query from query image and key from prompt source (proposed attention) instead of alternatives (ablation)SD-VICL · Pascal-5i foreground segmentationmIoU: 23.68 -> 55.49Improved
add AdaIN color alignment (ablation)SD-VICL · Pascal-5i foreground segmentationmIoU: 51.55 -> 55.49Improved
modify self-attention layers at all three resolutions instead of one (ablation)SD-VICL · Pascal-5i foreground segmentationmIoU: 50.33 -> 55.49Improved
use five prompts with 30 denoising steps instead of one prompt with 70 steps (ablation)SD-VICL · Pascal-5i foreground segmentationmIoU / inference time: 38.4 s (1 prompt, 70 steps) vs 35.9 s (5 prompts, 30 steps)Improved
increase number of prompts from 1 to 5 (ablation)SD-VICL · inference speedinference speedHurt
add text guidance to IMProv for keypoint detectionIMProv · DeepFashion keypoint detectionMSE: 31.30 -> 44.67Hurt
use contrast strength above 1.0 and enable swap-guidance (ablation)SD-VICL · Pascal-5i foreground segmentationmIoUImproved

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