Unsupervised Learning of Object Landmarks by Factorized Spatial Embeddings (2017-05-05T00:00:00.000000Z)

TL;DR

This paper proposes a novel unsupervised approach that can discover and learn landmarks in object categories, thus characterizing their structure, and shows that the learned landmarks establish meaningful correspondences between different object instances in a category without having to impose this requirement explicitly.

Abstract

Learning automatically the structure of object categories remains an important open problem in computer vision. In this paper, we propose a novel unsupervised approach that can discover and learn landmarks in object categories, thus characterizing their structure. Our approach is based on factorizing image deformations, as induced by a viewpoint change or an object deformation, by learning a deep neural network that detects landmarks consistently with such visual effects. Furthermore, we show that the learned landmarks establish meaningful correspondences between different object instances in a category without having to impose this requirement explicitly. We assess the method qualitatively on a variety of object types, natural and man-made. We also show that our unsupervised landmarks are highly predictive of manually-annotated landmarks in face benchmark datasets, and can be used to regress these with a high degree of accuracy.

Authors

A. Vedaldi

36 papers

Hakan Bilen

10 papers

James Thewlis

3 papers

TL;DR

Abstract

Authors

References84 items

AnchorNet: A Weakly Supervised Network to Learn Geometry-Sensitive Features for Semantic Matching

Convolutional Neural Network Architecture for Geometric Matching

Learning Features by Watching Objects Move

FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks

Robust Facial Landmark Detection via Recurrent Attentive-Refinement Networks

Fully-trainable deep matching

Learning Covariant Feature Detectors

Deep Deformation Network for Object Landmark Localization

Single Image 3D Interpreter Network

Context Encoders: Feature Learning by Inpainting

WarpNet: Weakly Supervised Matching for Single-View Reconstruction

Learning Dense Correspondence via 3D-Guided Cycle Consistency

Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzles

LIFT: Learned Invariant Feature Transform

Shuffle and Learn: Unsupervised Learning Using Temporal Order Verification

Colorful Image Colorization

Learning Representations for Automatic Colorization

300 Faces In-The-Wild Challenge: database and results

SSD: Single Shot MultiBox Detector

Part Localization using Multi-Proposal Consensus for Fine-Grained Categorization

Face alignment using cascade Gaussian process regression trees

FlowWeb: Joint image set alignment by weaving consistent, pixel-wise correspondences

Unsupervised learning of overcomplete face descriptors

Face alignment by coarse-to-fine shape searching

Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks

Unsupervised Visual Representation Learning by Context Prediction

Learning to See by Moving

FlowNet: Learning Optical Flow with Convolutional Networks

Adam: A Method for Stochastic Optimization

Deep Learning Face Attributes in the Wild

Facial Landmark Detection by Deep Multi-task Learning

Coarse-to-Fine Auto-Encoder Networks (CFAN) for Real-Time Face Alignment

Stacked Deformable Part Model with Shape Regression for Object Part Localization

Learning Deep Representation for Face Alignment with Auxiliary Attributes

One millisecond face alignment with an ensemble of regression trees

Face Alignment at 3000 FPS via Regressing Local Binary Features

Fine-Grained Visual Comparisons with Local Learning

A Compositional Model for Low-Dimensional Image Set Representation

Using a Deformation Field Model for Localizing Faces and Facial Points under Weak Supervision

Microsoft COCO: Common Objects in Context

DeepPose: Human Pose Estimation via Deep Neural Networks

Extensive Facial Landmark Localization with Coarse-to-Fine Convolutional Network Cascade

Guided Unsupervised Learning of Mode Specific Models for Facial Point Detection in the Wild

Bird Part Localization Using Exemplar-Based Models with Enforced Pose and Subcategory Consistency

Robust Face Landmark Estimation under Occlusion

Probabilistic Elastic Part Model for Unsupervised Face Detector Adaptation

Localizing Parts of Faces Using a Consensus of Exemplars

Robust Discriminative Response Map Fitting with Constrained Local Models

Deep Convolutional Network Cascade for Facial Point Detection

Learning to Align from Scratch

Representation Learning: A Review and New Perspectives

Face Alignment by Explicit Shape Regression

Real-time facial feature detection using conditional regression forests

Collection flow

Face detection, pose estimation, and landmark localization in the wild

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

Articulated pose estimation with flexible mixtures-of-parts

SIFT Flow: Dense Correspondence across Scenes and Its Applications

Object Detection with Discriminatively Trained Part Based Models

Cascaded pose regression

Facial point detection using boosted regression and graph models

Cat Head Detection - How to Effectively Exploit Shape and Texture Features

Automatic feature localisation with constrained local models

Extracting and composing robust features with denoising autoencoders

A Fast Learning Algorithm for Deep Belief Nets

Histograms of oriented gradients for human detection

Acquiring linear subspaces for face recognition under variable lighting

Active Appearance Models Revisited

Object class recognition by unsupervised scale-invariant learning

From Few to Many: Illumination Cone Models for Face Recognition under Variable Lighting and Pose

Towards automatic discovery of object categories

A Probabilistic Approach to Object Recognition Using Local Photometry and Global Geometry

Active Appearance Models

Categorization of faces using unsupervised feature extraction

Auto-association by multilayer perceptrons and singular value decomposition

Determining Optical Flow