AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

Reinforcing Generated Images via Meta-Learning for One-Shot Fine-Grained Visual Recognition

Satoshi Tsutsui, Yanwei Fu, David Crandall

TPAMI 2022 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Fine-grained and cross-domain few-shot baselines · 62 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 FinetuneGAN-generated images directly as augmentationProtoNet · CUB 5-way-1-shotaccuracy (%): 81.73 -> 79.40Hurt
add the learned image fusion module mixing real and generated images (ablation)ProtoNet · CUB 5-way-1-shotaccuracy (%): 81.73 -> 84.13Improved
add the learned image fusion module mixing real and generated images (ablation)ProtoNet · NAB 5-way-1-shotaccuracy (%): 87.91 -> 89.19Improved
add horizontal flip augmentationMetaIRNet (+FinetuneGAN) · CUB 5-way-1-shotaccuracy (%): 84.13 -> 84.80Improved
use FinetuneGAN images instead of FreezeDGAN images for mixing (ablation)MetaIRNet · CUB 5-way-1-shotaccuracy (%): 81.93 -> 84.13Improved
use FinetuneGAN images instead of jittered originals for mixing (ablation)MetaIRNet · CUB 5-way-1-shotaccuracy (%): 82.69 -> 84.13Improved
add hand-designed 3x3 binary mixed images to training dataOriginal (nearest neighbor) · CUB validation 5-way-1-shotaccuracy (%): 70.72 -> 71.50Improved
add plain generated images to training dataLogistic regression · CUB validation 5-way-1-shotaccuracy (%): 75.39 -> 74.08Hurt
add plain generated images to training dataSoftmax regression · CUB validation 5-way-1-shotaccuracy (%): 74.61 -> 73.55Hurt
add hand-designed mixed images to training dataSoftmax regression · CUB validation 5-way-1-shotaccuracy (%): 74.61 -> 75.40Improved
use 5x5 mixing weights instead of 3x3 (ablation)MetaIRNet · CUB 5-way-1-shotaccuracy (%): 84.13 -> 83.63Hurt
increase number of generated images n_aug from 1 to 5 (ablation)MetaIRNet · CUB 5-way-1-shotaccuracy (%)No clear effect

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