Image Clustering On Stl 10

Accuracy

Backbone

평가 결과

이 벤치마크에서 각 모델의 성능 결과

			Paper Title
TURTLE (CLIP + DINOv2)	0.997	-	Let Go of Your Labels with Unsupervised Transfer
TEMI DINO ViT-B	0.985	ViT-B	Exploring the Limits of Deep Image Clustering using Pretrained Models
TAC	0.982	-	Image Clustering with External Guidance
SPICE-BPA	0.943	ResNet-34	The Balanced-Pairwise-Affinities Feature Transform
DPAC	0.934	ResNet-34	Deep Online Probability Aggregation Clustering
SPICE*	0.929	ResNet-34	SPICE: Semantic Pseudo-labeling for Image Clustering
HUME	0.908	ResNet-18	-
TCL	0.868	ResNet-34	Twin Contrastive Learning for Online Clustering
RUC	0.867	ResNet-18	Improving Unsupervised Image Clustering With Robust Learning
IMC-SwAV (Best)	0.853	ResNet-18	Information Maximization Clustering via Multi-View Self-Labelling
CC	0.85	ResNet34	Contrastive Clustering
SeCu	0.836	ResNet-18	Stable Cluster Discrimination for Deep Clustering
IMC-SwAV (Avg+-)	0.831	ResNet-18	Information Maximization Clustering via Multi-View Self-Labelling
ITAE	0.8276	ViT-B/14	Improving Image Clustering with Artifacts Attenuation via Inference-Time Attention Engineering
SCAN	0.809	ResNet-18	SCAN: Learning to Classify Images without Labels
SCAN (Avg)	0.767	ResNet-18	SCAN: Learning to Classify Images without Labels
IDFD	0.756	ResNet-18	Clustering-friendly Representation Learning via Instance Discrimination and Feature Decorrelation
MiCE	0.752	ResNet-34	MiCE: Mixture of Contrastive Experts for Unsupervised Image Clustering
ConCURL	0.749	-	Representation Learning for Clustering via Building Consensus
MMDC	0.694	ResNet18	Multi-Modal Deep Clustering: Unsupervised Partitioning of Images

0 of 29 row(s) selected.