Deep Spectral Methods: A Surprisingly Strong Baseline for Unsupervised Semantic Segmentation and Localization

TL;DR

This paper introduces a spectral segmentation approach inspired by traditional graph partitioning methods, which effectively decomposes images into meaningful segments and localizes objects without supervision, outperforming existing methods on complex datasets.

Contribution

The paper presents a novel spectral method for unsupervised segmentation and localization that leverages eigenvectors of feature affinity matrices from self-supervised networks, outperforming prior approaches.

Findings

Outperforms state-of-the-art in unsupervised segmentation on Pascal VOC and MS-COCO.

Effectively localizes objects without labeled data.

Enables complex image editing tasks like background removal.

Abstract

Unsupervised localization and segmentation are long-standing computer vision challenges that involve decomposing an image into semantically-meaningful segments without any labeled data. These tasks are particularly interesting in an unsupervised setting due to the difficulty and cost of obtaining dense image annotations, but existing unsupervised approaches struggle with complex scenes containing multiple objects. Differently from existing methods, which are purely based on deep learning, we take inspiration from traditional spectral segmentation methods by reframing image decomposition as a graph partitioning problem. Specifically, we examine the eigenvectors of the Laplacian of a feature affinity matrix from self-supervised networks. We find that these eigenvectors already decompose an image into meaningful segments, and can be readily used to localize objects in a scene. Furthermore, by clustering the features associated with these segments across a dataset, we can obtain well-delineated, nameable regions, i.e. semantic segmentations. Experiments on complex datasets (Pascal VOC, MS-COCO) demonstrate that our simple spectral method outperforms the state-of-the-art in unsupervised localization and segmentation by a significant margin. Furthermore, our method can be readily used for a variety of complex image editing tasks, such as background removal and compositing.