AI Paper Atlas

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

Preserve Anything: Controllable Image Synthesis with Object Preservation

Prasen Kumar Sharma, Neeraj Matiyali, Siddharth Srivastava, Gaurav Sharma

ICCV 2025 · Image generation › Text-to-image diffusion · approach: Layout-to-image control · 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 aesthetic filtering of training data (ablation)RGB Only-BLIP2 · MS-COCO 3K evalFID: 17.31 -> 16.63Improved
add object mask condition (ablation)RGB Only-GPT4o · MS-COCO 3K evalFID: 19.37 -> 18.31Improved
add background guidance (ablation)RGB Only-GPT4o + Mask · MS-COCO 3K evalFID: 18.31 -> 15.26Improved
add background guidance (ablation)RGB Only-GPT4o + Mask · MS-COCO 3K evalLAION aesthetic: 6.11 -> 6.06Hurt
use GPT-4o captions instead of BLIP2 (ablation)RGB Only baseline · MS-COCO 3K evalCLIP-S: 29.68 -> 32.38Improved
use GPT-4o captions instead of BLIP2 (ablation)RGB Only baseline · MS-COCO 3K evalFID: 17.31 -> 19.37Hurt
switch backbone to SDXL (ablation)Preserve Anything · MS-COCO 3K evalFID: 15.26 -> 14.54Improved
switch backbone to SDXL (ablation)Preserve Anything · MS-COCO 3K evalCLIP-IQA: 0.6869 -> 0.6246Hurt
increase batch size from 16 to 128 (ablation)RGB Only-BLIP2 · MS-COCO 3K evalFID: 17.31 -> 16.98No clear effect
pretrain on 3D-rendered subset for lighting cues (ablation)Preserve Anything · GPT-4o critic of shadow/lightingcritic scoreImproved
use Preserve Anything instead of Inpaint AnythingPreserve Anything · User studyprompt alignment rating: 2.85 -> 3.58Improved
use Preserve Anything instead of Inpaint AnythingPreserve Anything · User studyAI artifacts rating (lower better): 2.91 -> 2.52Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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