AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

DICT: Data Injection and Contrastive Trajectory Refinement for Conditional Image Generation with Diffusion Models

Chunnan Shang, Zhizhong Wang, Xin Zhang, Hongwei Wang

ECCV 2026 · Image generation › Fast diffusion sampling · approach: Training-free diffusion acceleration · 0 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 Data Injection (noise-perturbed condition injected in early steps) (ablation)DICT · Text-driven style transferStyle Loss and CLIP LossImproved
add Data Injection (ablation)DICT · Text-driven style transferText ScoreNo clear effect
add Data Injection (ablation)DICT · FFHQ super-resolution and deblurringPSNR, SSIM, LPIPSImproved
add Contrastive Trajectory Refinement (ablation)DICT · Style transfer, super-resolution and deblurringquality metricsImproved
use DICT instead of TFG on text scoreDICT · Style transferText Score: 0.3092 -> 0.2952Hurt
use DICT instead of StyleShotDICT · Style transferStyle Loss: 0.6747 -> 0.6313Improved
use DICT instead of StyleShotDICT · Style transferCLIP Loss: 6.0415 -> 4.2334Improved
use DICT instead of FPS-SMCDICT · FFHQ super-resolutionPSNR: 27.6305 -> 28.8600Improved
use DICT instead of FPS-SMCDICT · FFHQ super-resolutionSSIM: 0.8283 -> 0.8233Hurt
use DICT instead of DCDPDICT · FFHQ deblurringPSNR: 27.9110 -> 27.5794Hurt
use DICT instead of DCDPDICT · FFHQ deblurringLPIPS: 0.2325 -> 0.2236Improved
vary Data Injection steps T1 too low or too high (ablation)DICT · Style transfervisual qualityMixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 5 of 5 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