Deep Spectral Methods: A Surprisingly Strong Baseline for Unsupervised Semantic Segmentation and Localization.
Luke Melas-KyriaziChristian RupprechtIro LainaAndrea VedaldiPublished in: CoRR (2022)
Keyphrases
- semantic segmentation
- spectral methods
- weakly supervised
- manifold learning
- random walk
- superpixels
- spectral clustering
- conditional random fields
- semi supervised
- unsupervised learning
- object categories
- topic models
- object classes
- machine learning
- multi label
- high dimensional
- data sets
- co occurrence
- supervised learning
- multiscale
- feature extraction
- decision trees
- computer vision
- learning algorithm