Semi-Supervised Image Classification

Semi-supervised image classification leverages unlabelled data as well as labelled data to increase classification performance.

You may want to read some blog posts to get an overview before reading the papers and checking the leaderboards:

An overview of proxy-label approaches for semi-supervised learning - Sebastian Ruder
Semi-Supervised Learning in Computer Vision - Amit Chaudhary

( Image credit: Self-Supervised Semi-Supervised Learning )

Papers

Recently Added Most Hyped Most Active Needs Verification Most Verified

Showing 151–167 of 167 papers

Title	Date	Tasks	Status
Interpolation Consistency Training for Semi-Supervised Learning	Mar 9, 2019	General ClassificationSemi-Supervised Image Classification	CodeCode Available
Semi-Supervised Deep Learning with Memory	Sep 1, 2018	Deep LearningGeneral Classification	CodeCode Available
Two Methods for Domain Adaptation of Bilingual Tasks: Delightfully Simple and Broadly Applicable	Jul 1, 2018	Bilingual Lexicon InductionClassification	CodeCode Available
There Are Many Consistent Explanations of Unlabeled Data: Why You Should Average	Jun 14, 2018	Domain AdaptationSemi-Supervised Image Classification	CodeCode Available
Unsupervised Learning using Pretrained CNN and Associative Memory Bank	May 2, 2018	Few-Shot Image ClassificationFine-Grained Image Classification	—Unverified
CapsuleGAN: Generative Adversarial Capsule Network	Feb 17, 2018	General ClassificationGenerative Adversarial Network	CodeCode Available
Structured Generative Adversarial Networks	Nov 2, 2017	image-classificationImage Classification	CodeCode Available
Triangle Generative Adversarial Networks	Sep 19, 2017	AttributeGenerative Adversarial Network	CodeCode Available
Good Semi-supervised Learning that Requires a Bad GAN	May 27, 2017	General ClassificationSemi-Supervised Image Classification	CodeCode Available
Temporal Ensembling for Semi-Supervised Learning	Oct 7, 2016	General ClassificationSemi-Supervised Image Classification	CodeCode Available
Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learning	Jun 14, 2016	Data AugmentationSemi-Supervised Image Classification	—Unverified
Unsupervised High-level Feature Learning by Ensemble Projection for Semi-supervised Image Classification and Image Clustering	Feb 2, 2016	ClassificationClustering	—Unverified
Learning The Structure of Deep Convolutional Networks	Dec 1, 2015	image-classificationImage Classification	—Unverified
Stacked What-Where Auto-encoders	Jun 8, 2015	DecoderImage Classification	CodeCode Available
CURL: Co-trained Unsupervised Representation Learning for Image Classification	May 29, 2015	ClassificationGeneral Classification	—Unverified
10,000+ Times Accelerated Robust Subset Selection (ARSS)	Sep 12, 2014	10-shot image generationAction Recognition	—Unverified
Comment on "Ensemble Projection for Semi-supervised Image Classification"	Aug 29, 2014	ClassificationGeneral Classification	—Unverified

Show:10 25 50

← PrevPage 4 of 4Next →

All datasets ImageNet - 10% labeled data ImageNet - 1% labeled data CIFAR-10, 4000 Labels cifar-100, 10000 Labels CIFAR-10, 250 Labels SVHN, 1000 labels SVHN, 250 Labels CIFAR-10, 40 Labels CIFAR-100, 400 Labels CIFAR-10, 1000 Labels CIFAR-100, 2500 Labels STL-10, 1000 Labels

Benchmark Results

#	Model	Metric	Claimed	Verified	Status
1	SimCLR (ResNet-50 4×)	Top 5 Accuracy	92.6	—	Unverified
2	Rotation + VAT + Ent. Min.	Top 5 Accuracy	91.23	—	Unverified
3	SimCLR (ResNet-50 2×)	Top 5 Accuracy	91.2	—	Unverified
4	Mean Teacher (ResNeXt-152)	Top 5 Accuracy	90.89	—	Unverified
5	OBoW (ResNet-50)	Top 5 Accuracy	90.7	—	Unverified
6	R2-D2 (ResNet-18)	Top 5 Accuracy	90.48	—	Unverified
7	FixMatch	Top 5 Accuracy	89.13	—	Unverified
8	UDA	Top 5 Accuracy	88.52	—	Unverified
9	SimCLR (ResNet-50)	Top 5 Accuracy	87.8	—	Unverified
10	DHO (ViT-Large)	Top 1 Accuracy	85.9	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	DHO (ViT-Large)	Top 1 Accuracy	84.6	—	Unverified
2	OBoW (ResNet-50)	Top 5 Accuracy	82.9	—	Unverified
3	DHO (ViT-Base)	Top 1 Accuracy	81.6	—	Unverified
4	REACT (ViT-Large)	Top 1 Accuracy	81.6	—	Unverified
5	Semi-SST (ViT-Huge)	Top 1 Accuracy	80.7	—	Unverified
6	Meta Co-Training	Top 1 Accuracy	80.7	—	Unverified
7	Super-SST (ViT-Huge)	Top 1 Accuracy	80.3	—	Unverified
8	Semi-ViT (ViT-Huge)	Top 1 Accuracy	80	—	Unverified
9	Semi-ViT (ViT-Large)	Top 1 Accuracy	77.3	—	Unverified
10	Super-SST (ViT-Small distilled)	Top 1 Accuracy	76.9	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	Γ-model	Percentage error	20.4	—	Unverified
2	GAN	Percentage error	15.59	—	Unverified
3	Bad GAN	Percentage error	14.41	—	Unverified
4	Triple-GAN-V2 (CNN-13, no aug)	Percentage error	12.41	—	Unverified
5	Pi Model	Percentage error	12.16	—	Unverified
6	SESEMI SSL (ConvNet)	Percentage error	11.65	—	Unverified
7	VAT	Percentage error	11.36	—	Unverified
8	GLOT-DR	Percentage error	10.6	—	Unverified
9	VAT+EntMin	Percentage error	10.55	—	Unverified
10	Triple-GAN-V2 (CNN-13)	Percentage error	10.01	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	Ⅱ-Model	Percentage error	39.19	—	Unverified
2	SESEMI SSL (ConvNet)	Percentage error	38.7	—	Unverified
3	Temporal ensembling	Percentage error	38.65	—	Unverified
4	R2-D2 (CNN-13)	Percentage error	32.87	—	Unverified
5	Dual Student (480)	Percentage error	32.77	—	Unverified
6	UPS (CNN-13)	Percentage error	32	—	Unverified
7	SHOT-VAE	Percentage error	25.3	—	Unverified
8	LiDAM	Percentage error	23.22	—	Unverified
9	EnAET (WRN-28-2-Large)	Percentage error	22.92	—	Unverified
10	FixMatch (RA, WRN-28-8)	Percentage error	22.6	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	Ⅱ-Model	Percentage error	53.12	—	Unverified
2	MixUp	Percentage error	47.43	—	Unverified
3	MeanTeacher	Percentage error	47.32	—	Unverified
4	VAT	Percentage error	36.03	—	Unverified
5	LiDAM	Percentage error	19.17	—	Unverified
6	MixMatch	Percentage error	11.08	—	Unverified
7	RealMix	Percentage error	9.79	—	Unverified
8	EnAET	Percentage error	7.6	—	Unverified
9	ReMixMatch	Percentage error	6.27	—	Unverified
10	FixMatch+CR	Percentage error	5.04	—	Unverified