Papers › ViCE: Improving Dense Representation Learning by Superpixelization and Contrasting...

ViCE: Improving Dense Representation Learning by Superpixelization and Contrasting Cluster Assignment

24 Nov 2021arXiv:2111.12460archive 2025-07-28

Robin Karlsson, Tomoki Hayashi, Keisuke Fujii, Alexander Carballo, Kento Ohtani, Kazuya Takeda

Recent self-supervised models have demonstrated equal or better performance than supervised methods, opening for AI systems to learn visual representations from practically unlimited data. However, these methods are typically classification-based and thus ineffective for learning high-resolution feature maps that preserve precise spatial information. This work introduces superpixels to improve self-supervised learning of dense semantically rich visual concept embeddings. Decomposing images into a small set of visually coherent regions reduces the computational complexity by 𝒪(1000) while preserving detail. We experimentally show that contrasting over regions improves the effectiveness of contrastive learning methods, extends their applicability to high-resolution images, improves overclustering performance, superpixels are better than grids, and regional masking improves performance. The expressiveness of our dense embeddings is demonstrated by improving the SOTA unsupervised semantic segmentation benchmark on Cityscapes, and for convolutional models on COCO.

Paper PDF Code

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

robin-karlsson0/vice officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Contrastive Learning Domain Generalization Learning Word Embeddings Representation Learning Self-Supervised Learning Semantic Segmentation Superpixels Unsupervised Semantic Segmentation Word Embeddings

Results from the paper archive 2025-07-28

Task	Dataset	Model	Metric	Value	Rank at snapshot	Leaderboard	Report
Unsupervised Semantic Segmentation	COCO-Stuff-27	ViCE	Clustering [Accuracy]	64.8	#20 of 29	Archive leaderboard	report
Unsupervised Semantic Segmentation	COCO-Stuff-27	ViCE	Clustering [mIoU]	21.77	#20 of 29	Archive leaderboard	report
Unsupervised Semantic Segmentation	Cityscapes test	ViCE	Accuracy	84.3	#2 of 14	Archive leaderboard	report
Unsupervised Semantic Segmentation	Cityscapes test	ViCE	mIoU	25.2	#2 of 14	Archive leaderboard	report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

1x1 Convolution Average Pooling Batch Normalization Bottleneck Residual Block Contrastive Learning Convolution Global Average Pooling Kaiming Initialization Max Pooling ReLU Residual Block Residual Connection SuperpixelGridMasks

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections