Temporal Alignment of Human Motion

Temporal alignment of three subjects kicking a ball.

These three sequences are recorded with different sensors (top row video, middle row motion capture and bottom row accelerometers).

People

Introduction

Temporal alignment of human motion has been a topic of recent interest due to its applications in animation, tele-rehabilitation and activity recognition among others. This paper presents canonical time warping (CTW)[1][2][3], an extension of dynamic time warping (DTW)[4] for temporally aligning multi-modal sequences from multiple subjects performing similar activities. CTW solves three major drawbacks of existing approaches based on DTW:

Code

Available at here.

Result

Temporal alignment of three multi-modal sequences by GTW.

The first two sequences are taken from CMU Motion Capture Database and Weizmann Action Database. The last sequence records the hand movement by an accelerometer. You can reproduce the same result using the function demoMix.m in the code.

Download the [Video 6MB].

Temporal alignment of two motion capture sequences by CTW.

The motion capture sequences are taken from CMU Grand Chanllenge Dataset. You can reproduce the same result using the function demoKit.m in the code.

Download the [Video 3MB].

Publications

References