Few Shot Image Classification On Cub 200 5 1

评估指标

Accuracy

评测结果

各个模型在此基准测试上的表现结果

		Paper Title	Repository
PT+MAP+SF+BPA (transductive)	95.80	The Balanced-Pairwise-Affinities Feature Transform
PT+MAP+SF+SOT (transductive)	95.80	The Self-Optimal-Transport Feature Transform
CAML [Laion-2b]	95.8	Context-Aware Meta-Learning
PT+MAP+SF (transductive)	95.48	Few-Shot Learning by Integrating Spatial and Frequency Representation
PEMnE-BMS*	94.78	Squeezing Backbone Feature Distributions to the Max for Efficient Few-Shot Learning
Illumination Augmentation	94.73	Sill-Net: Feature Augmentation with Separated Illumination Representation
LST+MAP	91.68	Transfer learning based few-shot classification using optimal transport mapping from preprocessed latent space of backbone neural network
PT+MAP	91.55%	Leveraging the Feature Distribution in Transfer-based Few-Shot Learning
EASY 3xResNet12 (transductive)	90.56	EASY: Ensemble Augmented-Shot Y-shaped Learning: State-Of-The-Art Few-Shot Classification with Simple Ingredients
EASY 4xResNet12 (transductive)	90.5	EASY: Ensemble Augmented-Shot Y-shaped Learning: State-Of-The-Art Few-Shot Classification with Simple Ingredients
BAVARDAGE	90.42	Adaptive Dimension Reduction and Variational Inference for Transductive Few-Shot Classification	-
ICI	89.58	Instance Credibility Inference for Few-Shot Learning
Transfer+SGC	88.35%	Graph-based Interpolation of Feature Vectors for Accurate Few-Shot Classification
ESPT	85.45	ESPT: A Self-Supervised Episodic Spatial Pretext Task for Improving Few-Shot Learning
TDM	84.36	Task Discrepancy Maximization for Fine-grained Few-Shot Classification
BD-CSPN + ESFR (ResNet-18)	82.68	Unsupervised Embedding Adaptation via Early-Stage Feature Reconstruction for Few-Shot Classification
TIM-GD	82.2%	Transductive Information Maximization For Few-Shot Learning
LaplacianShot	80.96	Laplacian Regularized Few-Shot Learning	-
S2M2R	80.68	Charting the Right Manifold: Manifold Mixup for Few-shot Learning
RENet	79.49	Relational Embedding for Few-Shot Classification

0 of 36 row(s) selected.