Learning Feature Pyramids for Human Pose Estimation

Abstract

Articulated human pose estimation is a fundamental yet challenging task in computer vision. The difficulty is particularly pronounced in scale variations of human body parts when camera view changes or severe foreshortening happens. Although pyramid methods are widely used to handle scale changes at inference time, learning feature pyramids in deep convolutional neural networks (DCNNs) is still not well explored. In this work, we design a Pyramid Residual Module (PRMs) to enhance the invariance in scales of DCNNs. Given input features, the PRMs learn convolutional filters on various scales of input features, which are obtained with different subsampling ratios in a multi-branch network. Moreover, we observe that it is inappropriate to adopt existing methods to initialize the weights of multi-branch networks, which achieve superior performance than plain networks in many tasks recently. Therefore, we provide theoretic derivation to extend the current weight initialization scheme to multi-branch network structures. We investigate our method on two standard benchmarks for human pose estimation. Our approach obtains state-of-the-art results on both benchmarks. Code is available at https://github.com/bearpaw/PyraNet.

Submitted to ICCV 2017

References in corpus (19)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Deep Residual Learning for Image Recognition
• DeepPose: Human Pose Estimation via Deep Neural Networks
• Wide Residual Networks
• Multi-Scale Context Aggregation by Dilated Convolutions
• Going Deeper with Convolutions
• Rethinking the Inception Architecture for Computer Vision
• Human pose estimation via Convolutional Part Heatmap Regression
• Stacked Hourglass Networks for Human Pose Estimation
• Aggregated Residual Transformations for Deep Neural Networks
• Fractional Max-Pooling
• Convolutional Pose Machines
• Pose Invariant Embedding for Deep Person Re-identification
• Multi-Context Attention for Human Pose Estimation
• Single-Image Depth Perception in the Wild
• Inside-Outside Net: Detecting Objects in Context with Skip Pooling and Recurrent Neural Networks
• DeepCut: Joint Subset Partition and Labeling for Multi Person Pose Estimation
• Human Pose Estimation using Deep Consensus Voting
• Zoom Out-and-In Network with Recursive Training for Object Proposal

Cited by in corpus (28)

• Human Pose Regression by Combining Indirect Part Detection and Contextual Information
• Numerical Coordinate Regression with Convolutional Neural Networks
• Simple Baselines for Human Pose Estimation and Tracking
• Analyzing Human-Human Interactions: A Survey
• FishNet: A Versatile Backbone for Image, Region, and Pixel Level Prediction
• Efficient Convolutional Neural Networks for Depth-Based Multi-Person Pose Estimation
• 3D Human Pose Estimation in the Wild by Adversarial Learning
• Optical Flow Guided Feature: A Fast and Robust Motion Representation for Video Action Recognition
• Look at Boundary: A Boundary-Aware Face Alignment Algorithm
• Multi-Scale Structure-Aware Network for Human Pose Estimation
• Person Re-identification with Deep Similarity-Guided Graph Neural Network
• Integral Human Pose Regression
• Pose2Seg: Detection Free Human Instance Segmentation
• Quantized Densely Connected U-Nets for Efficient Landmark Localization
• LightTrack: A Generic Framework for Online Top-Down Human Pose Tracking
• Learning to Refine Human Pose Estimation
• 3D Human Pose Estimation with 2D Marginal Heatmaps
• A Context-and-Spatial Aware Network for Multi-Person Pose Estimation
• Human Motion Analysis with Deep Metric Learning
• EfficientHRNet: Efficient Scaling for Lightweight High-Resolution Multi-Person Pose Estimation
• Matrix and tensor decompositions for training binary neural networks
• Ordinal Depth Supervision for 3D Human Pose Estimation
• StarMap for Category-Agnostic Keypoint and Viewpoint Estimation
• Deep Feature Pyramid Reconfiguration for Object Detection
• Bi-directional Graph Structure Information Model for Multi-Person Pose Estimation
• Key Frame Proposal Network for Efficient Pose Estimation in Videos
• Skeleton Transformer Networks: 3D Human Pose and Skinned Mesh from Single RGB Image
• Computer Vision and Abnormal Patient Gait Assessment a Comparison of Machine Learning Models