BROWSE
Detail
Conference Paper
Video-based Measurement of Physiological Parameters Using Peak-to-Valley Method for Minimization of Initial Dead Zone
Cited 0 time in 
Download 4 time
Share 
Download 4 time
Share
- Authors
- Youngjae Lim, Jang-Hee Yoo
- Issue Date
- 2018-12
- Citation
- International Conference on Machine Learning and Data Engineering (iCMLDE) 2018, pp.1-6
- Language
- English
- Type
- Conference Paper
- Project Code
- 18HS1700, Basic Technology for Extracting High-level Information from Multiple Sources Data base on Intelligent Analysis, Yoo Jang-Hee
- Abstract
- In this paper, we propose a new method to improve the accuracy and real-Time performance in iPPG (image photoplethysmography) approach which measures physiological parameters such as heart rate and oxygen saturation using PC camera. Most conventional iPPG methods extract physiological parameters by frequency or spectral analysis of signal traces in the moving temporal sliding window with fixed size. In this case, in order to minimize the influence of noise due to fluctuate in reflected light or face pose, the window size is made large enough in general, but it causes dead zone as much as the window size at the initial measurement time. In this study, we use size-increasing sliding window and modified peak-To-valley method instead of frequency analysis methods with fixed size window such as FFT or wavelet transform. Including these concepts, we have implemented a remote vital sign measurement system based on the real-Time pipeline architecture which tracks the ROIs, extracts appropriate signal traces from the ROIs, pre-/post-processes to remove noise, and applies frequency/spectral analysis. Finally, the performance of the proposed method was verified by comparing measurement data from the proposed system and reference data from the contact pulse oximeter using RMSE and Bland-Altman plot evaluation methods.
- KSP Keywords
- Analysis method, Bland-Altman plot, Dead Zone, Evaluation method, Face pose, Heart rate, Physiological parameters, Pulse Oximeter, Real-time performance, Reference data, Reflected light