Paper Info
Title
Serial Fusion Of Eulerian And Lagrangian Approaches For Accurate Heart-Rate Estimation Using Face Videos
Abstract
Camera-equipped devices are ubiquitous and proliferating in the day-to-day life. Accurate heart rate (HR) estimation from the face videos acquired from the low cost cameras in a non-contact manner, can be used in many real-world scenarios and hence, require rigorous exploration. This paper has presented an accurate and near real-time HR estimation system using these face videos. It is based on the phenomenon that the color and motion variations in the face video are closely related to the heart beat. The variations also contain the noise due to facial expressions, respiration, eye blinking and environmental factors which are handled by the proposed system. Neither Eulerian nor Lagrangian temporal signals can provide accurate HR in all the cases. The cases where Eulerian temporal signals perform spuriously are determined using a novel poorness measure and then both the Eulerian and Lagrangian temporal signals are employed for better HR estimation. Such a fusion is referred as serial fusion. Experimental results reveal that the error introduced in the proposed algorithm is 1.8 +/- 3.6 which is significantly lower than the existing well known systems.
Year
DOI
Venue
2017
10.1109/EMBC.2017.8037447
2017 39TH ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY (EMBC)
Field
DocType
Volume
Computer vision,Lagrangian,Heart beat,Computer science,Fusion,Eulerian path,Facial expression,Eye blinking,Artificial intelligence
Conference
2017
ISSN
Citations 
PageRank 
1094-687X
0
0.34
References 
Authors
6
3
Name
Order
Citations
PageRank
Puneet Gupta11158117.59
Brojeshwar Bhowmick2369.60
Arpan Pal319551.41