AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

GenHancer: Imperfect Generative Models are Secretly Strong Vision-Centric Enhancers

Shijie Ma, Yuying Ge, Teng Wang, Yuxin Guo, Yixiao Ge, Ying Shan

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Interleaved image-text generation with MLLMs · 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
enhance CLIP with GenHancer instead of DIVA (lightweight denoiser)OpenAI ViT-L-14 @224 · MMVP-VLMaverage accuracy: 25.9 -> 31.9Improved
enhance CLIP with GenHancer instead of DIVASigLIP ViT-SO-14 @224 · MMVP-VLMaverage accuracy: 40.7 -> 42.2Improved
use enhanced CLIP in LLaVA-1.5 versus original CLIPVicuna-7B · MMVP-MLLMaccuracy: 24.7 -> 30.7Improved
use enhanced CLIP in LLaVA-1.5 versus original CLIPVicuna-7B · CV-Bench 3Daccuracy: 58.7 -> 63.2Improved
use GenHancer CLIP versus DIVA CLIP in LLaVA-1.5Vicuna-7B · MMVP-MLLMaccuracy: 31.3 -> 30.7Hurt
post-tune CLIP with GenHancer on global semanticsOpenAI ViT-L-14 @224 · ImageNet-1K zero-shotaccuracy: 75.5 -> 75.6No clear effect
use scaled Logit-Normal timestamp sampling (s=1) instead of uniform (ablation)continuous denoiser · MMVP-VLM OpenAI@224average accuracy: 21.5 -> 31.9Improved
use too large scale s=10 in timestamp sampling instead of s=1 (ablation)continuous denoiser · MMVP-VLM OpenAI@224average accuracy: 31.9 -> 20.7Hurt
use two-stage training instead of end-to-end training (ablation)GenHancer · MMVP-VLMaverage accuracyImproved
condition only on the class token instead of adding local tokens (ablation)GenHancer · MMVP-VLMaverage accuracyImproved
use LoRA instead of direct training of CLIP in stage 2 (ablation)GenHancer · MMVP-VLMaverage accuracyImproved
use lightweight denoiser instead of pre-trained FLUX.1-dev (ablation)OpenAI ViT-L-14 @224 · MMVP-VLMaverage accuracy: 32.6 -> 31.9No clear effect

Later papers that compared with this method

1 later accepted paper reports GenHancer as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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