Image segmentation without training labels by clustering features with STEGO

Image segmentation without training labels by clustering features with STEGO

Unsupervised Semantic Segmentation by Distilling Feature Correspondences

arXiv paper abstract https://arxiv.org/abs/2203.08414

arXiv PDF paper https://arxiv.org/pdf/2203.08414.pdf

Unsupervised semantic segmentation aims to discover and localize semantically meaningful categories within image corpora without any form of annotation.

... algorithms must produce features for every pixel that are both semantically meaningful and compact enough to form distinct clusters.

... show that current unsupervised feature learning frameworks already generate dense features whose correlations are semantically consistent.

... motivates ... design STEGO (Self-supervised Transformer with Energy-based Graph Optimization) ... that distills unsupervised features into high-quality discrete semantic labels.

... STEGO ... encourages features to form compact clusters while preserving their relationships across the corpora.

STEGO ... significant improvement over the prior state of the art, on ... CocoStuff (+14 mIoU) and Cityscapes (+9 mIoU) semantic segmentation challenges.

Please like and share this post if you enjoyed it using the buttons at the bottom! Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact Web site with my other posts by category https://morrislee1234.wixsite.com/website LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b #ComputerVision #Segmentation #AINewsClips #AI #ML #ArtificialIntelligence #MachineLearning