AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Learn2Synth: Learning Optimal Data Synthesis Using Hypergradients for Brain Image Segmentation

Xiaoling Hu, Xiangrui Zeng, Oula Puonti, Juan Eugenio Iglesias, Bruce Fischl, Yaël Balbastre

ICCV 2025 · Detection and segmentation › Open-set and few-shot detection · approach: Interactive and referring segmentation · 5 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
learn augmentation parameters with hypergradients (Learn2Synth, parametric fixed sigma)Naive SynthSeg · OASIS3 (2D)Dice: 0.831 -> 0.860Improved
learn augmentation parameters with hypergradients (Learn2Synth, parametric fixed sigma)Naive SynthSeg · ABIDE (2D)Dice: 0.869 -> 0.878Improved
use nonparametric UNet augmentation (fixed sigma) instead of naive synthesisNaive SynthSeg · OASIS3 (2D)Dice: 0.831 -> 0.875Improved
Learn2Synth instead of mixing real images into SynthSeg trainingMixed SynthSeg · OASIS3 (2D)Dice: 0.854 -> 0.875Improved
fine-tune SynthSeg on real images versus Learn2SynthFinetuned SynthSeg · ABIDE (2D)Dice: 0.871 -> 0.874No clear effect
replace accurate parametric noise+INU model with nonparametric model (ablation)Learn2Synth noise+INU · synthetic noise + INU, train sigma=0 test sigma=0Dice: 0.924 -> 0.903Hurt
use nnUNet instead of UNet backbone for Learn2Synth (ablation)Learn2Synth (parametric, fixed sigma) · ABIDE (2D)Dice: 0.878 -> 0.893Improved
use model selection by validation set for supervised UNet (ablation)Supervised UNet (29 train) · FLASH 3degDice: 0.301 -> 0.419Improved
use validation set for model selection in Learn2Synth (ablation)Learn2Synth (29 train) · FLASH 3degDice: 0.801 -> 0.804No clear effect
reduce Learn2Synth training set from 29 to 1 sample (ablation)Learn2Synth · Buckner39 test (MPRAGE)Dice: 0.895 -> 0.718Hurt
Learn2Synth instead of SynthSeg on unseen FLASH contrastSynthSeg · FLASH 5degDice: 0.694 -> 0.789Improved
Learn2Synth instead of supervised UNet on unseen FLASH contrastSupervised UNet · FLASH 3degDice: 0.419 -> 0.804Improved

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