Object Segmentation Without Labels with Large-Scale Generative Models (2020-06-08T00:00:00.000000Z)

TL;DR

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, achieving new state-of-the-art results.

Abstract

The recent rise of unsupervised and self-supervised learning has dramatically reduced the dependency on labeled data, providing effective image representations for transfer to downstream vision tasks. Furthermore, recent works 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 standard benchmarks, we outperform existing unsupervised alternatives for object segmentation, achieving new state-of-the-art.

Authors

Artem Babenko

5 papers

S. Morozov

2 papers

A. Voynov

1 papers

TL;DR

Abstract

Authors

References38 items

Few-Cost Salient Object Detection with Adversarial-Paced Learning

Self-Supervised Learning for Large-Scale Unsupervised Image Clustering

Generative Pretraining From Pixels

Unsupervised Learning of Visual Features by Contrasting Cluster Assignments

Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning

Learning To Classify Images Without Labels

A Simple Framework for Contrastive Learning of Visual Representations

Unsupervised Discovery of Interpretable Directions in the GAN Latent Space

OneGAN: Simultaneous Unsupervised Learning of Conditional Image Generation, Foreground Segmentation, and Fine-Grained Clustering

Momentum Contrast for Unsupervised Visual Representation Learning

Large Scale Adversarial Representation Learning

Unsupervised Object Segmentation by Redrawing

Emergence of Object Segmentation in Perturbed Generative Models

Salient Object Detection in the Deep Learning Era: An In-Depth Survey

Video Saliency Detection Using Object Proposals

Large Scale GAN Training for High Fidelity Natural Image Synthesis

Deep Unsupervised Saliency Detection: A Multiple Noisy Labeling Perspective

Co-Saliency Detection for RGBD Images Based on Multi-Constraint Feature Matching and Cross Label Propagation

Supervision by Fusion: Towards Unsupervised Learning of Deep Salient Object Detector

Learning to Detect Salient Objects with Image-Level Supervision

Salient Object Detection via Structured Matrix Decomposition

Correspondence Driven Saliency Transfer

The Cityscapes Dataset for Semantic Urban Scene Understanding

Hierarchical Image Saliency Detection on Extended CSSD

A weighted sparse coding framework for saliency detection

U-Net: Convolutional Networks for Biomedical Image Segmentation

Saliency Optimization from Robust Background Detection

Salient Object Detection: A Discriminative Regional Feature Integration Approach

Saliency Detection via Graph-Based Manifold Ranking

Hierarchical Saliency Detection

Geodesic Saliency Using Background Priors

Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials

Global contrast based salient region detection

SUN database: Large-scale scene recognition from abbey to zoo

ImageNet: A large-scale hierarchical image database

DeepUSPS: Deep Robust Unsupervised Saliency Prediction via Self-supervision

The caltech-ucsd birds

Delving into the Whorl of Flower Segmentation

Field of Study

Venue Information

Name

Type

URL

Alternate Names