Algorithm To Calculate Pulse from PPG Signal After Eliminating Touch Errors from the Fingertip Video Captured by Smartphone Camera
Authors:
Ayan Chatterjee,
Sundar Gopalakrishnan,
Martin Gerdes,
Santiago Martinez,
Nibedita Pahari,
Pankaj Khatiwada
Abstract:
With the ongoing heart problems of the population worldwide, the medical requirements of the people are expected to increase. Electrocardiogram (ECG) is one of the proven to capture the heart response signal to assess the electrical and muscular functions of the heart. The ECG setup is expensive and needs proper training, and of course, it is not instant. For fast, accurate heart parameter monitor…
▽ More
With the ongoing heart problems of the population worldwide, the medical requirements of the people are expected to increase. Electrocardiogram (ECG) is one of the proven to capture the heart response signal to assess the electrical and muscular functions of the heart. The ECG setup is expensive and needs proper training, and of course, it is not instant. For fast, accurate heart parameter monitoring, scientists pay attention to the photoplethysmogram signal (PPG), based on the light intensity of a particular wavelength. Android smartphone with a good quality camera has come to ordinary people's reach and has become one of the most necessary and rugged devices for today and future generations. We can use its powerful features to solve or assess heart state monitoring by capturing the image's necessary data. The mobile camera has a photo emitting diode and a photodetector. The light source illuminates the tissue. The photodetector calculates the small variation in light intensity associated with blood volume change in the vessels (mainly fingertips, toes, and ears). We have captured unfocused contact video to capture PPG using an Android Smartphone. Then, we removed a certain percent of camera touch errors based on average pixel intensity count in the red plane, and it is a new approach that has been introduced in this research. We used a 2nd order Butterworth (IIR) band pass filter for noise removal, FFT Hann Window for frequency analysis and leakage reduction. We have developed an algorithm using MATLAB as a development platform, for accurate pulse (BPM) measurement. Moreover, we have done a comparative analysis of developed algorithm with other available algorithms for PPG-based pulse calculation. In this study, the fingertip video was captured when the body was at rest
△ Less
Submitted 6 September, 2021; v1 submitted 30 November, 2020;
originally announced December 2020.
Automated Human Activity Recognition by Colliding Bodies Optimization-based Optimal Feature Selection with Recurrent Neural Network
Authors:
Pankaj Khatiwada,
Ayan Chatterjee,
Matrika Subedi
Abstract:
In smart healthcare, Human Activity Recognition (HAR) is considered to be an efficient model in pervasive computation from sensor readings. The Ambient Assisted Living (AAL) in the home or community helps the people in providing independent care and enhanced living quality. However, many AAL models were restricted using many factors that include computational cost and system complexity. Moreover,…
▽ More
In smart healthcare, Human Activity Recognition (HAR) is considered to be an efficient model in pervasive computation from sensor readings. The Ambient Assisted Living (AAL) in the home or community helps the people in providing independent care and enhanced living quality. However, many AAL models were restricted using many factors that include computational cost and system complexity. Moreover, the HAR concept has more relevance because of its applications. Hence, this paper tempts to implement the HAR system using deep learning with the data collected from smart sensors that are publicly available in the UC Irvine Machine Learning Repository (UCI). The proposed model involves three processes: (1) Data collection, (b) Optimal feature selection, (c) Recognition. The data gathered from the benchmark repository is initially subjected to optimal feature selection that helps to select the most significant features. The proposed optimal feature selection is based on a new meta-heuristic algorithm called Colliding Bodies Optimization (CBO). An objective function derived by the recognition accuracy is used for accomplishing the optimal feature selection. Here, the deep learning model called Recurrent Neural Network (RNN) is used for activity recognition. The proposed model on the concerned benchmark dataset outperforms existing learning methods, providing high performance compared to the conventional models.
△ Less
Submitted 19 November, 2021; v1 submitted 7 October, 2020;
originally announced October 2020.
