AI Paper Atlas

Research map / Image generation / Editing and style transfer

DeepProtect: Proactive Face-Swapping Defense using Identity Blending and Attribute Distortion

Eungi Lee, Seung-hyeok Back, Hyung-Il Kim, Seok Bong Yoo

CVPR 2026 · Image generation › Editing and style transfer · approach: Identity-preserving face manipulation

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 generator optimization with identity-lock loss to identity blending (ablation)DeepProtect · CelebA-HQ, SimSwapSSIM of protected image: 0.584 -> 0.915Improved
add generator optimization when combining identity blending and attribute distortion (ablation)DeepProtect · CelebA-HQ, SimSwapPDS: -0.196 -> 0.192Improved
add identity blending to attribute distortion (ablation)DeepProtect · CelebA-HQ, SimSwapISM: 0.419 -> 0.201Improved
add attribute distortion to identity blending with generator optimization (ablation)DeepProtect · CelebA-HQ, SimSwapPDS: 0.136 -> 0.192Improved
use identity blending alone without generator optimization (ablation)DeepProtect · CelebA-HQ, SimSwapSSIM of protected image: 0.584Hurt
combine identity blending with attribute distortion instead of attribute distortion only (ablation)DeepProtect · CelebA-HQ, SimSwapDSR (%): 60.5 -> 94.8Improved
use DeepProtect Combined instead of FaceShieldDeepProtect · CelebA-HQ, SimSwapDSR (%): 81.6 -> 94.8Improved
apply LoRA to the affine transform module and tune only middle layersDeepProtect generator tuning · Complexitytrainable parameters: approximately 11M trainable parameters before LoRAImproved
apply Gaussian blur post-processing to protected imagesDeepProtect · CelebA-HQ, SimSwapDSR (%): 94.8 -> 93.5No clear effect
apply adaptive purification attackDeepProtect · CelebA-HQ, SimSwapDSR (%): 94.8 -> 93.2No clear effect
noise-based defenses under imperceptibility constraintCMUA-Watermark and DF-RAP · CelebA-HQ, SimSwapDSR (%): 41.0 and 42.7No clear effect

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