Timezone: »
The recent rise of unsupervised and self-supervised learning has dramatically reduced the dependency on labeled data, providing high-quality representations for transfer on downstream tasks. Furthermore, recent works also employed these representations in a fully unsupervised setup for image classification, reducing the need for human labels on the fine-tuning stage as well. This work demonstrates that large-scale unsupervised models can also perform a more challenging object segmentation task, requiring neither pixel-level nor image-level labeling. Namely, we show that recent unsupervised GANs allow to differentiate between foreground/background pixels, providing high-quality saliency masks. By extensive comparison on common benchmarks, we outperform existing unsupervised alternatives for object segmentation, achieving new state-of-the-art.
Author Information
Andrey Voynov (Yandex)
Stanislav Morozov (Yandex)
Artem Babenko (Yandex)
Related Events (a corresponding poster, oral, or spotlight)
-
2021 Spotlight: Object Segmentation Without Labels with Large-Scale Generative Models »
Fri. Jul 23rd 02:25 -- 02:30 AM Room
More from the Same Authors
-
2021 : Discovering Weight Initializers with Meta Learning »
Dmitry Baranchuk · Artem Babenko -
2023 Poster: TabDDPM: Modelling Tabular Data with Diffusion Models »
Akim Kotelnikov · Dmitry Baranchuk · Ivan Rubachev · Artem Babenko -
2021 Poster: Functional Space Analysis of Local GAN Convergence »
Valentin Khrulkov · Artem Babenko · Ivan Oseledets -
2021 Spotlight: Functional Space Analysis of Local GAN Convergence »
Valentin Khrulkov · Artem Babenko · Ivan Oseledets -
2020 Poster: Unsupervised Discovery of Interpretable Directions in the GAN Latent Space »
Andrey Voynov · Artem Babenko -
2019 Poster: Learning to Route in Similarity Graphs »
Dmitry Baranchuk · Dmitry Persiyanov · Anton Sinitsin · Artem Babenko -
2019 Oral: Learning to Route in Similarity Graphs »
Dmitry Baranchuk · Dmitry Persiyanov · Anton Sinitsin · Artem Babenko