Human Pose Forecasting via Deep Markov Models

Abstract

Human pose forecasting is an important problem in computer vision with applications to human-robot interaction, visual surveillance, and autonomous driving. Usually, forecasting algorithms use 3D skeleton sequences and are trained to forecast for a few milliseconds into the future. Long-range forecasting is challenging due to the difficulty of estimating how long a person continues an activity. To this end, our contributions are threefold: (i) we propose a generative framework for poses using variational autoencoders based on Deep Markov Models (DMMs); (ii) we evaluate our pose forecasts using a pose-based action classifier, which we argue better reflects the subjective quality of pose forecasts than distance in coordinate space; (iii) last, for evaluation of the new model, we introduce a 480,000-frame video dataset called Ikea Furniture Assembly (Ikea FA), which depicts humans repeatedly assembling and disassembling furniture. We demonstrate promising results for our approach on both Ikea FA and the existing NTU RGB+D dataset.

Accepted to DICTA'17

References in corpus (11)

• DeepPose: Human Pose Estimation via Deep Neural Networks
• Stochastic Backpropagation and Approximate Inference in Deep Generative Models
• Tutorial on Variational Autoencoders
• Generating Videos with Scene Dynamics
• Stacked Hourglass Networks for Human Pose Estimation
• Convolutional Pose Machines
• Unsupervised Learning for Physical Interaction through Video Prediction
• Learning Stochastic Recurrent Networks
• Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields
• Flowing ConvNets for Human Pose Estimation in Videos
• Recurrent Network Models for Human Dynamics