Image Clustering On Imagenet Dog 15

评估指标

ARI

Accuracy

Backbone

NMI

评测结果

各个模型在此基准测试上的表现结果

模型名称	ARI	Accuracy	Backbone	NMI	Paper Title	Repository
DPAC	0.598	0.726	ResNet-34	0.667	Deep Online Probability Aggregation Clustering	-
C3	0.28	0.434	-	0.448	C3: Cross-instance guided Contrastive Clustering	-
DAC	-	0.275	-	0.219	Deep Adaptive Image Clustering
DCCM	-	0.383	-	0.321	Deep Comprehensive Correlation Mining for Image Clustering	-
MiCE	0.286	0.439	-	0.423	MiCE: Mixture of Contrastive Experts for Unsupervised Image Clustering	-
DEC	-	0.195	-	0.122	Unsupervised Deep Embedding for Clustering Analysis	-
MAE-CT (best)	0.879	0.943	ViT-H/16	0.904	Contrastive Tuning: A Little Help to Make Masked Autoencoders Forget	-
CC	0.274	0.429	-	0.445	Contrastive Clustering	-
ConCURL	0.531	0.695	-	0.63	Representation Learning for Clustering via Building Consensus	-
TCL	0.516	0.644	-	0.623	Twin Contrastive Learning for Online Clustering	-
ProPos*	0.675	0.775	ResNet-34	0.737	Learning Representation for Clustering via Prototype Scattering and Positive Sampling	-
IDFD	0.413	0.591	-	0.546	Clustering-friendly Representation Learning via Instance Discrimination and Feature Decorrelation	-
VAE	-	0.179	-	0.107	Auto-Encoding Variational Bayes	-
PRO-DSC	-	0.840	-	0.812	Exploring a Principled Framework For Deep Subspace Clustering
GAN	-	0.174	-	0.121	Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks	-
SPICE	0.526	0.675	-	0.627	SPICE: Semantic Pseudo-labeling for Image Clustering	-
CoHiClust	0.232	0.355	ResNet-50	0.411	Contrastive Hierarchical Clustering	-
JULE	-	0.138	-	0.054	Joint Unsupervised Learning of Deep Representations and Image Clusters	-
MAE-CT (mean)	0.821	0.874	ViT-H/16	0.882	Contrastive Tuning: A Little Help to Make Masked Autoencoders Forget	-
ProPos	0.627	0.745	ResNet-34	0.692	Learning Representation for Clustering via Prototype Scattering and Positive Sampling	-

0 of 20 row(s) selected.