HE ZHAO

129 Hallock Road , Lake Grove, NY 11755
(631)-786-0768 (cell), (631)-444-2354 (office), he.zhao@gmail.com

I'm looking for a job ...
If you have a position available, please contact me. I will be really appreciated about that. Thanks!

Algorithms Research

1. Extracting respiratory rate from photoplethysmographic (PPG) signal
2. Development of time-varying surrogate data (Paper in preparation)
3. A new parametric method to estimate coherence function (Paper in preparation)
4. Time-varying Causal Coherence Function and its Application to Renal Blood Pressure and Blood Flow Data (Published journal paper)
5. Estimation of time-varying coherence function using time-varying transfer functions (Published journal paper)
6. ARMA model coefficients identification: using multiple Basis Function Sets (Publised journal paper)

Recent Projects

1. Biomedical data acquisition and on-board real-time signal processing
2. Wireless multiple vital parameters monitoring system (Published conference paper)

Past Projects

1. Photoplethysmographic (PPG) signal acquisition circuit
2 . Virtual endoscopy (Published journal paper)
3. LFP901 simulation system (part-time)

Publications

======================================================================================================================================

Extracting respiratory rate from photoplethysmographic (PPG) signal (top)
	A new method has been developed to extract respiratory rate from PPG signal. With this technique, not only hear rate can be extracted from PPG signal, but respiratory rate is also obtainable. The significance of extracting respiratory rate from PPG signal has been fully addressed in the literatures. Since PPG is a non-invasive monitoring method, it causes much interests in the application of intensive care unit (ICU) and infant vital parameter monitoring. Extracting respiratory rate from PPG signal eliminates the requirement of an extract respiratory sensor, thus reduces the monitoring cost and simplifies the monitoring system. In many commercially available devices, independent component analysis (ICA) method is widely used to extract respiratory rate from PPG. However, the ICA method requires both the red and infrared waveforms to extract respiratory rate. Other method, for example, digital filtering method, although only needs one channel waveform, but is subjected to limited respiratory frequency band and requires a priori knowledge of the respiratory rate. The method presented here only needs one channel waveform and does not require any a priori knowledge of the respiratory rate. To test this method, a series of experiments have been conducted. Healthy subjects were required to breathe at some fixed frequencies both in supine and upright positions. The preliminary results show that the performance of this method is similar regardless of postures. For all the subjects, this method is able to extract the respiratory rate almost all the time. However, In order to precisely quantify the performance of this method, more experiments need to be administrated. The selected results shown below are results at different respiratory rate of 0.1Hz, 0.2Hz, 0.3Hz, 0.4Hz and 0.5Hz. For each trial, 3 minutes of PPG is recorded at a sampling rate of 200 Hz. The resiratory rate is calculated every 2 seconds. In the figures shown below, the upper panels are the recorded PPG signal; the middle panels are the extracted respiratory rate; and the bottom panels are the hear rate.

	Subject breathes at 0.1 Hz (click to see larger figure)
	Subject breathes at 0.2 Hz.(click to see larger figure)
	Subject breathes at 0.3 Hz.(click to see larger figure)
	Subject breathes at 0.4 Hz.(click to see larger figure)
	Subject breathes at 0.5Hz. (click to see larger figure)

Biomedical data acquisition and on-board real-time signal processing (top)
	Achievements in this project
		>> Integrated analog electrocardiograph (ECG) circuit and blood oxygen module (MP506 by Nellcor) with MSP430F449 microcontroller >> Developed an embedded software platform that is capable of data acquisition, data transmission and real-time signal processing
	Project description
		This project is to develop an independent module that performs real-time digital signal processsing (DSP). This module will be integrated into the wireless multiple vital parameters monitoring system, which is described below. The demand for on-board DSP is significant since the device runs on batteries. Power consumption will be greatly reduced if only critical physiology parameters are transmitted instead of the whole raw data. For example, transmitting one lead raw ECG signal at 200Hz sampling rate, the Tmote Sky consumes approximately 25 mA current. To make a comparison, we implemented a real-time R-wave detection algorithm at the device end to calculate heart rate. For every 5 minutes, the power spectrum density (PSD) of heart rate is calculated and the ratio between low frequency (LF) and high frequency (HF) is computed and sent to the monitoring center. In this scenario, the average power consumption of Tmote Sky is approximately only 4 mA. Although, the on-board DSP program may increase the microcontroller’s power consumption a little bit, it dramatically saves the power consumed by radio transmission. Another great advantage of on-board DSP is that it also saves a lot of bandwith. In the tests aforementioned, the transmission loads are calculated as the following, respectively: raw data transmition: 2 byte/point × 200 points/sec = 400 bytes/sec after on-board DSP: 2 byte/parameter × 1 parameter/5 min < 0.007 bytes/sec The calculation shows that bandwidth is dramatically saved by transmitting only critical parameters.
	Development tools
		>> IAR Embedded Workbench >> Embedded c++ >> TI MSP430F449 microcontorller

Wireless multiple vital parameters monitoring system (top)

Achievements in this project

>> Integrated the Tmote Sky wireless transmission module with an ECG circuit and SpO2 module (MP506 by Nellcor)
>> Developed an application layer communication protocol Command/Data Transmission Protocol (C/D TP) tailored for this system
>> Developed the embedded software both at the transmitter and receiver ends
>> Developed the display software at the display terminate, which features data recording, real-time signal processing, and internet data transmission

Project description

The goal of this project is to develop a portable, low-cost, and battery-powered wireless monitoring system that is capable of measuring multiple physiological parameters simultaneously from many subjects. This device features in on-command variable signal sampling rates and is capable of performing limited digital signal processing on-board. With these features, only a few derived physiological parameters are transmitted in normal conditions, which saves both power consumption and radio bandwidth. Once abnormality is detected in the parameters, the health carer at the remote care center can request the transmission of raw signal by simply issuing a command to the device wirelessly. The wireless communication of data is based on a commercially-available mote known as Tmote Sky. The star network topology (SNT) is used to collect data from many patients via multiple motes. Based on the standard specifications of the mote, the SNT strategy, and the C/D TP protocol, a single mote can support up to 10 electrocardiogram signals with a sampling rate of 200 Hz for each. A pulse oximeter (MP506) has also been incorporated in the device to provide SpO2 readings. This capability facilitates affordable wireless monitoring of multiple physiologic signals from many subjects; its application is especially attractive for monitoring subjects in nursing homes, battlefields, and disaster scenarios.

Results: hardware modules

Tmote Sky wireless module

ECG circuit

SpO2 module (MP506, Nellcor)

Integrated ECG transmitter (left) and receiver (right)

Results: display software snapshots (click to see larger images)

The display software was developed using Delphi. Currently, the displaysoftware runs on a common personal computer. Its features include:
1. Getting data from the receiver via USB port, or via Internet, and showing the data in real-time fashion (ECG signal and SpO2 readings);
2. Performing real-time R-wave detection on ECG signal to extract heart rate data;
3. Performing real-time FFT on extracted heart rate data and calculating frequency statistics;
4. Relaying data to another display software runing on another computer by Internet.

Main interface of display software

Channel selection and system confiuration window of the display software


Raw ECG signal is displayed in real-time fashion	Real-time Power Spectrum Density of Heart Rate data is calcualted and shown in real-time. The statistics in frequency domain such as the power in low frequency (LF), high frequency (HF) and very high frequency (VLH) are also shown, together with the raios among them.	Real-time Time-frequency spectrum and instantaneous hear rate and blood oxygen readings are updated very one second

Photoplethysmographic (PPG) signal acquisition circuit (top)
	When red/infrared lights penetrate the blood vessels, the lights will be absorbed and scattered by oxyhaemoglobin (HbO2). The blood oxygen saturation can be measured based on the ratio between the absorption of red/infrared lights. The built circuit is to detect and amplify the attenuated lights of red/infrared. Figures shown below are components and schematic diagrams of the circuit. (click to see larger figures)
	The system schematic diagram :
	The LEDs	The photodiode

	The amplified waveforms (red and infrared channel)

Virtual endoscopy (top)

Achievements in this project

>> Developed a new camera model for Virtual Endoscopy System;
>> Successfully implemented the model, and promoted the reconstruction speed;
>> Revised the PHONG illumination model;
>> Raised a "Dichotomy Resample with Varied Step" method, greatly increased the reconstruction speed;
>> Director and programmer of this project.

Project description

Virtual Endoscope (VE) is the combination of Medical Image Visualization and Virtual Reality (VR). It emerged only several years ago. Although it has not developed completely, it has shown its great promising future. Compared with traditional Fiber Endoscope (FE), VE has some unique characters. It is a non-invasive examination method, which can reduce the patient’s pains and lowers the risk of bleeding and infection. It can reach the parts for examination that FE cannot reach, providing more diagnosis information that the physician may need. It is also a good candidate for General Medical Examination and can partly replace the FE to save the cost of medicine. At the same time, VE plays an important role in the diagnosis of tumors, surgery simulation and surgery planning, providing more detailed anatomic information. It also can be used for the countercheck method after surgery. Besides, VE is also an excellent implement for teaching and helping the patient to understand his illness. The technique of VE is also an important part in the medical robot. If VE can be combined with micromechanism and robot technique, it will bring the medical robots into our real life.

Results

Still images

	Human intestine	Human Intestine

	Human Spinal Bone	Human Spinal Bone

	Human Stomach	Human Stomach
	Movie: Human trachea virtual tour

LFP901 simulation system (part-time) (top)

Role in this project

>> System designer and programmer of the project;
>> Director of the project.

Project description

This is a pc-based simulation software system. This software is designed for training. As we all know, electric system is under strict monitoring and protection for safety purpose. The monitoring and protection job is taken by some specially designed systems. The device LFP901 is a major part of the monitoring and protection system. However, professional training is needed for the operators that are in charge of the monitoring and protection system, despite the system is of high automation. This software is developed for the training, or teaching and learning purpose. One can set different exception in the software, and observe what happens to the system, what the reponses of the system. At the same time,the simulation software also issues different warning signals, such as light, sound ect., just exactly like the real system does.
Since the software is for training purpose, when the simulation is started, it records every operation that has been performed. The benifit is obvious: when the simulation is over, one can review his performance during the simulation and see which step he did wrong.
This project is my part-time job when I was in the frist year of my graduate study. The goal of the project was targeted the market of Electric Simulation Softwares. For financial reasons, together with others, this project stopped on half of the way to its goal.

Snapshots of software interface


Main Interface	Operation Record	Exception Setup Interface

CZX11a Simulation Interface	LFP901a Simulation Interface	LFX912 Simulation Interface

Setting Printing Interface	Trip Printing Interface

Publications (top)
	Zhao H. ; Cupples W.A. ; Ju K. ; Chon K. H. “Time-varying Causal Coherence Function and its Application to Renal Blood Pressure and Blood Flow Data”, IEEE Trans. Biomed. Eng. (In press) Xinnian Chen, Irene Solomon, He Zhao, Ki Chon, "Fast Oscillatory Rhythms in Inspiratory Motor Discharge: A Mathematical Model", Proceeding of EMBC'06, Aug. 30-Sept.3, 2006. New York, USA He Zhao, Xinnian Chen, Ki Hwan Ju, Ernst Raeder, Ki H. Chon, "A Portable, Low-cost, Battery-powered Wireless Monitoring System for Obtaining Multiple Physiological Parameters from Multiple Subjects", Proceeding of EMBC’06, Aug. 30-Sept. 3, 2006. New York, USA H. Zhao, S. Lu, R. Zou, K. Ju, and K. H. Chon, "Estimation of time-varying coherence function using time-varying transfer functions," Ann Biomed Eng, vol. 33, pp. 1582-94, 2005. Zhao H, Ju K, Chon KH. “An approach to estimate time-varying casual coherence function.” Methods Inf. Med. 2007;46(2):102-9. K. H. Chon, H. Zhao, R. Zou and K. Ju. Multiple Time-Varying Dynamic Analyses Using Multiple Sets of Basis Functions. IEEE Transaction on Biomedical Engineering, Vol 52, No. 5, pp 956-960, May 2005. YE Zhu; ZHAO He; FENG Huan-qing, “Computation of Surface Normal Vector in Virtual Endoscope System”, Journal of Biomedical Engineering Research, vol. 22 (4): 12—15, 2003. Tao MEI, Qinghua HUANG, Heqin ZHOU, He ZHAO and Huanqing FENG, “An Improved Multiscale Image Enhancement via Laplacian Pyramid”, The 2nd International Conference on Image and Graphics, SPIE vol.4875, pp402-410, Aug.2002. Feiniu YUAN, Heqin ZHOU, He ZHAO, Huanqing FENG, “Sampled points decomposing based ray casting for virtual endoscopy”, The 2nd InternationalConference on Image and Graphics, SPIE vol.4875, Aug.2002. ZHAO He, ZHOU- Heqin, FENG Huanqing. “A PC-based Imaging Algorithm of Virtual Endoscope”. Space Medicine & Medical Engineering Feb. 2002, Vol.15 No.1: 59-63. ZHAO He, ZHOU Heqin, FENG Huanqing. Study on Illumination Model for Virtual Endoscope System. Proceedings of 2001 Annual National Conference on Biomedical Electronics, Biomedical Measurement, Biomedical Information & Control, Biomedical Sensor Techniques: 116—117, Wuhan China, 2001 ZHAO He, ZHOU Heqin, FENG Huanqing. Study on Imaging Algorithm for Virtual Endoscope System. Proceedings of 2001 Annual National Conference on Biomedical Electronics, Biomedical Measurement, Biomedical Information & Control, Biomedical Sensor Techniques: 289—290, Wuhan China, 2001 CONG Shuang, ZHAO He. “The Disadvantage and Improvement of Back-Propagation Network”, Automation Panorama, 1999.01: 25-27.