Browsing by Author "He, Da-Hong"

Now showing 1 - 2 of 2

A Region of Interest Selection for Vision-Based Respiratory Rate Measurement in Sleeping Scenario
(2020-01-01)
Wu, Bing-Fei
;
Lin, Chun-Han
;
Huang, Po-Wei
;
Chen, Kuan-Hung
;
He, Da-Hong
Vision-based respiratory measurement can remotely measure respiratory information without affecting the sleep quality of the user. However, several scenarios during sleeping, such as when blankets cover areas of the face or body, can be a challenge when estimate respiratory rate (RR) due to the uncertainty of the region of interest (ROI). In this study, we first investigate the metrics and physical meanings of ROI selection from the perspectives of array signal processing and the concept of linear combining. Then, we propose an ROI detection algorithm based on both temporal and spatial consistency (TSC), which aims to extract the representative characteristics of respiration with fewer computational resources. Furthermore, a sleeping database, containing more than 50 hours of data, was built to investigate the performance of benchmarked ROI methods during long-term sleeping. The experimental results demonstrate that TSC attains the highest signal-to-noise ratio (SNR) at 20.6 dB and a relatively low elapsed time of 100.2 frames per second (fps). For the RR estimation accuracy, TSC reduces the mean absolute error from 1.65 breaths per minute (bpm) to 0.97 bpm.
Remote Photoplethysmography Enhancement with Machine Leaning Methods
(2019-01-01)
Wu, Bing-Fei
;
Huang, Po-Wei
;
He, Da-Hong
;
Lin, Chung-Han
;
Chen, Kuan-Hung
Driver's physiological state is highly correlated to the traffic safety. An affordable and convenient way to monitor driver's physiological state is remote Photoplethysmography (rPPG). Earlier algorithms achieved high accuracy on measuring rPPG signals in stationary case. But in real cases, such as driving, rPPG signals might be corrupted with interference. To obtain higher Signal-to-Noise-Ratio (SNR) rPPG signals, three algorithms are proposed. The PCA spectral subtraction (PCA-SS) considers the spectrum of the environmental noise and utilizes the energy subtraction to reduce the noise. The machine learning methods, convolution autoencoder (CAE) and multi-channel convolution autoencoder (Multi-CAE), are adopted in order to enhance the rPPG signal. The test data we used are 187 videos recorded in stationary case, passenger case, and real driving situation. In driving situation, the Multi-CAE method, in comparison with the original method provided by W. Wang et al. [1] and G. De Haan et al. [2], achieves 33% & 35% reduction in MAE, RMSE respectively, and 11% improvement in success rate [3].