Unsupervised Semantic Segmentation by Distilling Feature Correspondences

Abstract

Unsupervised semantic segmentation aims to discover and localize semantically meaningful categories within image corpora without any form of annotation. To solve this task, algorithms must produce features for every pixel that are both semantically meaningful and compact enough to form distinct clusters. Unlike previous works which achieve this with a single end-to-end framework, we propose to separate feature learning from cluster compactification. Empirically, we show that current unsupervised feature learning frameworks already generate dense features whose correlations are semantically consistent. This observation motivates us to design STEGO ($\text{S}$elf-supervised $\text{T}$ransformer with $\text{E}$nergy-based $\text{G}$raph $\text{O}$ptimization), a novel framework that distills unsupervised features into high-quality discrete semantic labels. At the core of STEGO is a novel contrastive loss function that encourages features to form compact clusters while preserving their relationships across the corpora. STEGO yields a significant improvement over the prior state of the art, on both the CocoStuff ($\text{+14 mIoU}$) and Cityscapes ($\text{+9 mIoU}$) semantic segmentation challenges.

Cited by in corpus (6)

• DINOv2: Learning Robust Visual Features without Supervision
• Pixel-Level Clustering Network for Unsupervised Image Segmentation
• DatUS^2: Data-driven Unsupervised Semantic Segmentation with Pre-trained Self-supervised Vision Transformer
• Interpreting Vision and Language Generative Models with Semantic Visual Priors
• Leveraging Self-Supervised Vision Transformers for Segmentation-based Transfer Function Design
• SERE: Exploring Feature Self-relation for Self-supervised Transformer