Home Research Papers Datasets State of the Art Pricing

Discover, visualize, and connect AI research papers. Explore the latest trends and insights in artificial intelligence research.

Product

Home
Research Papers
About

Support

Contact
Terms of Service
Privacy Policy

© 2026 Papersgraph. All rights reserved.

computer-vision-10

Unsupervised Semantic Segmentation

3260 papers • 126 benchmarks • 313 datasets

Models that learn to segment each image (i.e. assign a class to every pixel) without seeing the ground truth labels. ( Image credit: SegSort: Segmentation by Discriminative Sorting of Segments )

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in unsupervised-semantic-segmentation-10

Trend

Dataset

Best Model

Actions

COCO-Stuff-27

COCO-Stuff-27

PASCAL VOC 2012 val

PASCAL VOC 2012 val

Cityscapes test

Libraries

i

Use these libraries to find unsupervised-semantic-segmentation-10 models and implementations

Datasets

Cityscapes

COCO-Stuff

Dark Zurich

ImageNet-S

SUIM

Nighttime Driving

Subtasks

Unsupervised Semantic Segmentation with Language-image Pre-training

Most implemented papers

Deep Clustering for Unsupervised Learning of Visual Features

Mathilde Caron, Piotr Bojanowski, Armand Joulin, Matthijs Douze•Sat Jul 14 2018

This work presents DeepCluster, a clustering method that jointly learns the parameters of a neural network and the cluster assignments of the resulting features and outperforms the current state of the art by a significant margin on all the standard benchmarks.

2126

Content

Introduction Benchmarks Datasets Subtasks Libraries Papers

Cityscapes test

COCO-Stuff-3

COCO-Stuff-3

Potsdam-3

Potsdam-3

ImageNet-S-50

ImageNet-S-50

COCO-Stuff-171

COCO-Stuff-171

COCO-Stuff-81

COCO-Stuff-81

COCO-Stuff-15

COCO-Stuff-15

COCO-Persons

COCO-Persons

Cityscapes val

Cityscapes val

Nighttime Driving

Nighttime Driving

Dark Zurich

Dark Zurich

ACDC (Adverse Conditions Dataset with Correspondences)

ACDC (Adverse Conditions Dataset with Correspondences)

COCO-All

COCO-All

ImageNet-S

ImageNet-S

ImageNet-S-300

ImageNet-S-300

SUIM

SUIM

Nighttime Driving

ACDC (Adverse Conditions Dataset with Correspondences)

ACDC (Adverse Conditions Dataset with Correspondences)

KITTI-STEP

0

Invariant Information Clustering for Unsupervised Image Classification and Segmentation

A. Vedaldi, João F. Henriques, Xu Ji•Mon Jul 16 2018

A novel clustering objective is presented that learns a neural network classifier from scratch, given only unlabelled data samples, and discovers clusters that accurately match semantic classes, achieving state-of-the-art results in eight unsupervised clustering benchmarks spanning image classification and segmentation.

923 0

Rectifying Pseudo Label Learning via Uncertainty Estimation for Domain Adaptive Semantic Segmentation

Zhedong Zheng, Yi Yang•Sat Mar 07 2020

The proposed approach dynamically sets different confidence thresholds according to the prediction variance, rectifies the learning from noisy pseudo labels, and achieves significant improvements over the conventional pseudo label learning and yields competitive performance on all three benchmarks.

562 0

Large-Scale Unsupervised Semantic Segmentation

Junwei Han, Philip H. S. Torr, Ming-Hsuan Yang, Mingg-Ming Cheng, Shangqi Gao, Zhong-Yu Li•Sat Jun 05 2021

This work proposes a new problem of large-scale unsupervised semantic segmentation (LUSS) with a newly created benchmark dataset to help the research progress and presents a simple yet effective method that works surprisingly well for LUSS.

118 0

Unsupervised Semantic Segmentation by Contrasting Object Mask Proposals

L. Gool, Simon Vandenhende, Wouter Van Gansbeke, Stamatios Georgoulis•Wed Feb 10 2021

A two-step framework that adopts a predetermined mid-level prior in a contrastive optimization objective to learn pixel embeddings and argues about the importance of having a prior that contains information about objects, or their parts, and discusses several possibilities to obtain such a prior in an unsupervised manner.

274 0

PiCIE: Unsupervised Semantic Segmentation using Invariance and Equivariance in Clustering

B. Hariharan, Kavita Bala, Jang Hyun Cho, Utkarsh Mall•Mon Mar 29 2021

The method, PiCIE (Pixel-level feature Clustering using Invariance and Equivariance), is the first method capable of segmenting both things and stuff categories without any hyperparameter tuning or task-specific pre-processing.

227 0

Unsupervised Semantic Segmentation by Distilling Feature Correspondences

Noah Snavely, W. Freeman, Bharath Hariharan, Zhoutong Zhang, Mark Hamilton•Tue Mar 15 2022

STEGO is a novel framework that distills unsupervised features into high-quality discrete semantic labels that encourages features to form compact clusters while preserving their relationships across the corpora.

300 0

ReCo: Retrieve and Co-segment for Zero-shot Transfer

Weidi Xie, Samuel Albanie, Gyungin Shin•Mon Jun 13 2022

This work leverages the retrieval abilities of one language-image pre-trained model, CLIP, to dynamically curate training sets from unlabelled images for arbitrary collections of concept names, and leverage the robust correspondences offered by modern image representations to co-segment entities among the resulting collections.

128 0

SILOP: An Automated Framework for Semantic Segmentation Using Image Labels Based on Object Perimeters

Erik Ostrowski, B. Prabakaran, Muhammad Shafique•Mon Mar 13 2023

An exhaustive analysis is conducted to illustrate that SILOP enhances existing state-of-the-art frameworks for image-level-based semantic segmentation, and proposes a framework that introduces an additional module using object perimeters for improved saliency.

1 0

Adding a benchmark result helps the community track progress.

Unsupervised Semantic Segmentation | State-of-the-Art