Estimating heart rate via depth video motion tracking

Cheng Yang, Gene Cheung, Vladimir Stankovic

Research output: Contribution to conferencePaper

11 Citations (Scopus)

Abstract

Depth sensors like Microsoft Kinect can acquire partial geometric
information in a 3D scene via captured depth images,
with potential application to non-contact health monitoring.
However, captured depth videos typically suffer from low
bit-depth representation and acquisition noise corruption, and
hence using them to deduce health metrics that require tracking
subtle 3D structural details is difficult. In this paper, we
propose to capture depth video using Kinect 2.0 to estimate
the heart rate of a human subject; as blood is pumped to circulate
through the head, tiny oscillatory head motion can be detected
for periodicity analysis. Specifically, we first perform a
joint bit-depth enhancement / denoising procedure to improve
the quality of the captured depth images, using a graph-signal
smoothness prior for regularization. We then track an automatically
detected nose region throughout the depth video to
deduce 3D motion vectors. The deduced 3D vectors are then
analyzed via principal component analysis to estimate heart
rate. Experimental results show improved tracking accuracy
using our proposed joint bit-depth enhancement / denoising
procedure, and estimated heart rates are close to ground truth.
LanguageEnglish
Number of pages6
Publication statusPublished - 3 Jul 2015
EventICME-2015 - Italy, Torino, Italy
Duration: 29 Jun 20153 Jul 2015

Conference

ConferenceICME-2015
CountryItaly
CityTorino
Period29/06/153/07/15

Fingerprint

Heart Rate
Head
Health
Periodicity
Principal Component Analysis
Nose
Principal component analysis
Noise
Blood
Joints
Monitoring
Sensors

Keywords

  • graph signal processing
  • health monitoring
  • image enhancement

Cite this

Yang, C., Cheung, G., & Stankovic, V. (2015). Estimating heart rate via depth video motion tracking. Paper presented at ICME-2015, Torino, Italy.
Yang, Cheng ; Cheung, Gene ; Stankovic, Vladimir. / Estimating heart rate via depth video motion tracking. Paper presented at ICME-2015, Torino, Italy.6 p.
@conference{6a36b2bedd9443faa2c2bd84af94c72e,
title = "Estimating heart rate via depth video motion tracking",
abstract = "Depth sensors like Microsoft Kinect can acquire partial geometricinformation in a 3D scene via captured depth images,with potential application to non-contact health monitoring.However, captured depth videos typically suffer from lowbit-depth representation and acquisition noise corruption, andhence using them to deduce health metrics that require trackingsubtle 3D structural details is difficult. In this paper, wepropose to capture depth video using Kinect 2.0 to estimatethe heart rate of a human subject; as blood is pumped to circulatethrough the head, tiny oscillatory head motion can be detectedfor periodicity analysis. Specifically, we first perform ajoint bit-depth enhancement / denoising procedure to improvethe quality of the captured depth images, using a graph-signalsmoothness prior for regularization. We then track an automaticallydetected nose region throughout the depth video todeduce 3D motion vectors. The deduced 3D vectors are thenanalyzed via principal component analysis to estimate heartrate. Experimental results show improved tracking accuracyusing our proposed joint bit-depth enhancement / denoisingprocedure, and estimated heart rates are close to ground truth.",
keywords = "graph signal processing, health monitoring, image enhancement",
author = "Cheng Yang and Gene Cheung and Vladimir Stankovic",
year = "2015",
month = "7",
day = "3",
language = "English",
note = "ICME-2015 ; Conference date: 29-06-2015 Through 03-07-2015",

}

Yang, C, Cheung, G & Stankovic, V 2015, 'Estimating heart rate via depth video motion tracking' Paper presented at ICME-2015, Torino, Italy, 29/06/15 - 3/07/15, .

Estimating heart rate via depth video motion tracking. / Yang, Cheng; Cheung, Gene; Stankovic, Vladimir.

2015. Paper presented at ICME-2015, Torino, Italy.

Research output: Contribution to conferencePaper

TY - CONF

T1 - Estimating heart rate via depth video motion tracking

AU - Yang, Cheng

AU - Cheung, Gene

AU - Stankovic, Vladimir

PY - 2015/7/3

Y1 - 2015/7/3

N2 - Depth sensors like Microsoft Kinect can acquire partial geometricinformation in a 3D scene via captured depth images,with potential application to non-contact health monitoring.However, captured depth videos typically suffer from lowbit-depth representation and acquisition noise corruption, andhence using them to deduce health metrics that require trackingsubtle 3D structural details is difficult. In this paper, wepropose to capture depth video using Kinect 2.0 to estimatethe heart rate of a human subject; as blood is pumped to circulatethrough the head, tiny oscillatory head motion can be detectedfor periodicity analysis. Specifically, we first perform ajoint bit-depth enhancement / denoising procedure to improvethe quality of the captured depth images, using a graph-signalsmoothness prior for regularization. We then track an automaticallydetected nose region throughout the depth video todeduce 3D motion vectors. The deduced 3D vectors are thenanalyzed via principal component analysis to estimate heartrate. Experimental results show improved tracking accuracyusing our proposed joint bit-depth enhancement / denoisingprocedure, and estimated heart rates are close to ground truth.

AB - Depth sensors like Microsoft Kinect can acquire partial geometricinformation in a 3D scene via captured depth images,with potential application to non-contact health monitoring.However, captured depth videos typically suffer from lowbit-depth representation and acquisition noise corruption, andhence using them to deduce health metrics that require trackingsubtle 3D structural details is difficult. In this paper, wepropose to capture depth video using Kinect 2.0 to estimatethe heart rate of a human subject; as blood is pumped to circulatethrough the head, tiny oscillatory head motion can be detectedfor periodicity analysis. Specifically, we first perform ajoint bit-depth enhancement / denoising procedure to improvethe quality of the captured depth images, using a graph-signalsmoothness prior for regularization. We then track an automaticallydetected nose region throughout the depth video todeduce 3D motion vectors. The deduced 3D vectors are thenanalyzed via principal component analysis to estimate heartrate. Experimental results show improved tracking accuracyusing our proposed joint bit-depth enhancement / denoisingprocedure, and estimated heart rates are close to ground truth.

KW - graph signal processing

KW - health monitoring

KW - image enhancement

UR - http://www.icme2015.ieee-icme.org/

M3 - Paper

ER -

Yang C, Cheung G, Stankovic V. Estimating heart rate via depth video motion tracking. 2015. Paper presented at ICME-2015, Torino, Italy.