SCOPS: Self-Supervised Co-Part Segmentation (2019-05-03T00:00:00.000000Z)

TL;DR

This work proposes a self-supervised deep learning approach for part segmentation, where several loss functions are devised that aids in predicting part segments that are geometrically concentrated, robust to object variations and are also semantically consistent across different object instances.

Abstract

Parts provide a good intermediate representation of objects that is robust with respect to camera, pose and appearance variations. Existing work on part segmentation is dominated by supervised approaches that rely on large amounts of manual annotations and also can not generalize to unseen object categories. We propose a self-supervised deep learning approach for part segmentation, where we devise several loss functions that aids in predicting part segments that are geometrically concentrated, robust to object variations and are also semantically consistent across different object instances. Extensive experiments on different types of image collections demonstrate that our approach can produce part segments that adhere to object boundaries and also more semantically consistent across object instances compared to existing self-supervised techniques.

Authors

Ming-Hsuan Yang

23 papers

Pavlo Molchanov

5 papers

V. Jampani

10 papers

TL;DR

Abstract

Authors

References57 items

Discovery of Latent 3D Keypoints via End-to-end Geometric Reasoning

Deep Feature Factorization For Concept Discovery

Deforming Autoencoders: Unsupervised Disentangling of Shape and Appearance

Unsupervised Discovery of Object Landmarks as Structural Representations

End-to-End Weakly-Supervised Semantic Alignment

Improving Landmark Localization with Semi-Supervised Learning

Unsupervised learning of object frames by dense equivariant image labelling

SCNet: Learning Semantic Correspondence

Unsupervised Learning of Object Landmarks by Factorized Spatial Embeddings

Hand Keypoint Detection in Single Images Using Multiview Bootstrapping

Network Dissection: Quantifying Interpretability of Deep Visual Representations

Proposal Flow: Semantic Correspondences from Object Proposals

Convolutional Neural Network Architecture for Geometric Matching

Semantic Co-segmentation in Videos

Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization

Do Semantic Parts Emerge in Convolutional Neural Networks?

DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs

Learning Deep Features for Discriminative Localization

Deep Residual Learning for Image Recognition

Stereo Matching by Training a Convolutional Neural Network to Compare Image Patches

Object-based RGBD image co-segmentation with mutex constraint

Fine-grained recognition without part annotations

The stitched puppet: A graphical model of 3D human shape and pose

Dense Semantic Correspondence Where Every Pixel is a Classifier

Learning to compare image patches via convolutional neural networks

Object Detectors Emerge in Deep Scene CNNs

Deep Learning Face Attributes in the Wild

Fully convolutional networks for semantic segmentation

How transferable are features in deep neural networks?

Computing the stereo matching cost with a convolutional neural network

Facial Landmark Detection by Deep Multi-task Learning

Human Pose Estimation with Fields of Parts

Very Deep Convolutional Networks for Large-Scale Image Recognition

Analyzing the Performance of Multilayer Neural Networks for Object Recognition

Saliency Optimization from Robust Background Detection

Detect What You Can: Detecting and Representing Objects Using Holistic Models and Body Parts

CNN Features Off-the-Shelf: An Astounding Baseline for Recognition

3D Object Representations for Fine-Grained Categorization

Visualizing and Understanding Convolutional Networks

Deformable Spatial Pyramid Matching for Fast Dense Correspondences

Unsupervised Joint Object Discovery and Segmentation in Internet Images

ImageNet classification with deep convolutional neural networks

Object Detection Using Strongly-Supervised Deformable Part Models

Unsupervised co-segmentation through region matching

Multi-class cosegmentation

Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials

Strong supervision from weak annotation: Interactive training of deformable part models

Annotated Facial Landmarks in the Wild: A large-scale, real-world database for facial landmark localization

The Caltech-UCSD Birds-200-2011 Dataset

From co-saliency to co-segmentation: An efficient and fully unsupervised energy minimization model

SIFT Flow: Dense Correspondence across Scenes and Its Applications

Poselets: Body part detectors trained using 3D human pose annotations

A discriminatively trained, multiscale, deformable part model

Learning the parts of objects by non-negative matrix factorization

Pietro Perona, and Serge Belongie. The caltech-ucsd birds

International Journal of Computer Vision manuscript No. (will be inserted by the editor) The PASCAL Visual Object Classes (VOC) Challenge

Deforming au-toencoders

Field of Study

Journal Information

Name

Page

Venue Information

Name

Type

URL

Alternate Names