MoDeep: A Deep Learning Framework Using Motion Features for Human Pose Estimation

TL;DR

This paper introduces MoDeep, a deep learning framework that leverages motion and color features for improved human pose estimation in videos, supported by a new dataset and outperforming existing methods.

Contribution

The paper presents a novel convolutional network architecture that integrates motion features and introduces the FLIC-motion dataset for enhanced pose estimation.

Findings

Significantly improved pose detection accuracy

Effective use of motion features in deep learning models

New dataset extends FLIC with motion information

Abstract

In this work, we propose a novel and efficient method for articulated human pose estimation in videos using a convolutional network architecture, which incorporates both color and motion features. We propose a new human body pose dataset, FLIC-motion, that extends the FLIC dataset with additional motion features. We apply our architecture to this dataset and report significantly better performance than current state-of-the-art pose detection systems.