AI Paper Atlas

Research map / Image generation / Editing and style transfer

NearID: Identity Representation Learning via Near-identity Distractors

Aleksandar Cvejic, Rameen Abdal, Abdelrahman Eldesokey, Bernard Ghanem, Peter Wonka

ECCV 2026 · Image generation › Editing and style transfer · approach: Diffusion face generation and editing · 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
train MAP head with NearID loss on frozen SigLIP2SigLIP2 · NearIDSSR (%): 30.74 -> 99.17Improved
train MAP head with NearID loss on frozen SigLIP2SigLIP2 · MTGSSR (%): 0.0 -> 35.0Improved
train MAP head with NearID loss on frozen SigLIP2SigLIP2 · DreamBench++human correlation: 0.516 -> 0.545Improved
use standard InfoNCE that ignores distractors (ablation)SigLIP2 MAP head · NearIDSSR (%): 30.74 -> 60.97Mixed
use Circle loss with ranking (ablation)SigLIP2 MAP head · DreamBench++human correlation: 0.516 -> 0.141Hurt
add oracle ranking to InfoNCE (ablation)InfoNCE · DreamBench++human correlation: 0.516 -> 0.167Hurt
use L_NearID instead of InfoNCE (ablation)SigLIP2 MAP head · DreamBench++human correlation: 0.555 -> 0.545Mixed
add positive cohesion term (ablation)L_NearID · NearID / MTGSSR (%): 99.17 -> 99.31; MTG 35.0 -> 36.0No clear effect
add NearID dataset to MTG-only training (ablation)NearID training data · NearIDSSR (%): from 99.2% to 58.3%Improved
add MTG data to NearID training (ablation)NearID training data · NearIDSSR (%): 0.8%Improved
use four inpainting engines instead of one (ablation)NearID training data · NearIDSSR (%): 98.4% -> 86.9%Improved
train on MTG only versus full pipeline (ablation)NearID training data · MTGSSR (%): 49.0% vs 35.0%Mixed

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