Tong Gao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Tong Gao has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Jingjing Zhang
School of Artificial Intelligence and Automation, China University of Geosciences, Wuhan, China
3 shared publications - Xuanhe Fan
Huazhong University of Science and Technology
2 shared publications - Zhijie Huang
China University of Geosciences
2 shared publications Fangwu Liu
Chinese Academy of Sciences
1 shared publication- Fangwu Liu
Shanghai Institute of Technical Physics, Chinese Academy of Sciences, 500 Yutian Road, Shanghai 200083, China
1 shared publication - Jia Fu
School of Automation, China University of Geosciences, Wuhan, China
1 shared publication
Latest Publications
- 8 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25 (10) , 16628-16639
Noncontact Video-Based Vital Signs Continuous Monitor for Unconstrained Conscious Rodents
Animal experiments should minimize pain and discomfort and maximize animal welfare to satisfy the experimental animal 3R principle. The quality of vital signs monitor results collected from conscious and stress-free rodents is superior, ensuring experimental data’s reliability, accuracy, and scientificity. This work proposes a noncontact video-based vital signs continuous monitoring method for unconstrained conscious rodents. First, rodents are tracked using keypoint detection, and regions of interest (ROI) on the rodents are selected based on signal-to-noise ratio (SNR) distribution of vital signs. Second, corner motion within the ROI region is tracked to synthesize skin pulsations, from which respiratory signals are extracted. Finally, a method is proposed to eliminate the influence of respiratory harmonics, thereby removing respiratory interference and extracting the heart rate (HR). Four rats and four mice were used to measure the performance of the proposed method. This work is the first to use a camera to continuously monitor the HR of unrestrained and conscious rodents. Compared with the ground truth [semi-implantable electrocardiogram (ECG)], the relative HR error was 1.86% $\pm ~0.82$ % and 1.20% $\pm ~0.33$ % for rats and mice, respectively. This method is used for unconstrained conscious rodents and has high applicability, allowing it to be used in various experimental environments and conditions. At the same time, this work contributes to the development of new methods for continuously and objectively evaluating animal welfare, thereby promoting the development of modern biomedical and ethical research.
- 8 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25 (10) , 16628-16639
Noncontact Video-Based Vital Signs Continuous Monitor for Unconstrained Conscious Rodents
Animal experiments should minimize pain and discomfort and maximize animal welfare to satisfy the experimental animal 3R principle. The quality of vital signs monitor results collected from conscious and stress-free rodents is superior, ensuring experimental data’s reliability, accuracy, and scientificity. This work proposes a noncontact video-based vital signs continuous monitoring method for unconstrained conscious rodents. First, rodents are tracked using keypoint detection, and regions of interest (ROI) on the rodents are selected based on signal-to-noise ratio (SNR) distribution of vital signs. Second, corner motion within the ROI region is tracked to synthesize skin pulsations, from which respiratory signals are extracted. Finally, a method is proposed to eliminate the influence of respiratory harmonics, thereby removing respiratory interference and extracting the heart rate (HR). Four rats and four mice were used to measure the performance of the proposed method. This work is the first to use a camera to continuously monitor the HR of unrestrained and conscious rodents. Compared with the ground truth [semi-implantable electrocardiogram (ECG)], the relative HR error was 1.86% $\pm ~0.82$ % and 1.20% $\pm ~0.33$ % for rats and mice, respectively. This method is used for unconstrained conscious rodents and has high applicability, allowing it to be used in various experimental environments and conditions. At the same time, this work contributes to the development of new methods for continuously and objectively evaluating animal welfare, thereby promoting the development of modern biomedical and ethical research.
- 1 March 2024
- Published by IOP Publishing in Physiological Measurement
- Vol. 45 (3) , 035005
Remote photoplethysmography based on reflected light angle estimation
Objective . In previous studies, the factors affecting the accuracy of imaging photoplethysmography (iPPG) heart rate (HR) measurement have been focused on the light intensity, facial reflection angle, and motion artifacts. However, the factor of specularly reflected light has not been studied in detail. We explored the effect of specularly reflected light on the accuracy of HR estimation and proposed an estimation method for the direction of specularly radiated light. Approach . To study the HR measurement accuracy influenced by specularly reflected light, we control the component of specularly reflected light by controlling its angle. A total of 100 videos from four different reflected light angles were collected, and 25 subjects participated in the dataset collection. We extracted angles and illuminations for 71 facial regions, fitting sample points through interpolation, and selecting the angle corresponding to the maximum weight in the fitted curve as the estimated reflected angle. Main results . The experimental results show that higher specularly reflected light compromises HR estimation accuracy under the same value of light intensity. Notably, at a 60° angle, the HR accuracy (ACC) increased by 0.7%, while the signal-to-noise ratio and Pearson correlation coefficient increased by 0.8 dB and 0.035, respectively, compared to 0°. The overall root mean squared error, standard deviation, and mean error of our proposed reflected light angle estimation method on the illumination multi-angle incidence (IMAI) dataset are 1.173°, 0.978°, and 0.773°. The average Pearson value is 0.8 in the PURE rotation dataset. In addition, the average ACC of HR measurements in the PURE dataset is improved by 1.73% in our method compared to the state-of-the-art traditional methods. Significance . Our method has great potential for clinical applications, especially in bright light environments such as during surgery, to improve accuracy and monitor blood volume changes in blood vessels.
- 1 March 2024
- Published by IOP Publishing in Physiological Measurement
- Vol. 45 (3) , 035005
Remote photoplethysmography based on reflected light angle estimation
Objective . In previous studies, the factors affecting the accuracy of imaging photoplethysmography (iPPG) heart rate (HR) measurement have been focused on the light intensity, facial reflection angle, and motion artifacts. However, the factor of specularly reflected light has not been studied in detail. We explored the effect of specularly reflected light on the accuracy of HR estimation and proposed an estimation method for the direction of specularly radiated light. Approach . To study the HR measurement accuracy influenced by specularly reflected light, we control the component of specularly reflected light by controlling its angle. A total of 100 videos from four different reflected light angles were collected, and 25 subjects participated in the dataset collection. We extracted angles and illuminations for 71 facial regions, fitting sample points through interpolation, and selecting the angle corresponding to the maximum weight in the fitted curve as the estimated reflected angle. Main results . The experimental results show that higher specularly reflected light compromises HR estimation accuracy under the same value of light intensity. Notably, at a 60° angle, the HR accuracy (ACC) increased by 0.7%, while the signal-to-noise ratio and Pearson correlation coefficient increased by 0.8 dB and 0.035, respectively, compared to 0°. The overall root mean squared error, standard deviation, and mean error of our proposed reflected light angle estimation method on the illumination multi-angle incidence (IMAI) dataset are 1.173°, 0.978°, and 0.773°. The average Pearson value is 0.8 in the PURE rotation dataset. In addition, the average ACC of HR measurements in the PURE dataset is improved by 1.73% in our method compared to the state-of-the-art traditional methods. Significance . Our method has great potential for clinical applications, especially in bright light environments such as during surgery, to improve accuracy and monitor blood volume changes in blood vessels.
- 13 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72 (00189456) , 1-12
Robust Heart Rate Measurement by Adaptive ROI Selection for Head-Rotation Conditions
Region of interest (ROI) selection is a key process for remote heart rate measurement approach. Previous studies have shown that the selected ROI could easily be lost under head rotation circumstances and does not perform well under drastic changes in illumination. To solve the problems, this work proposed a novel adaptive ROI selection method. Firstly, we collected a dataset for modeling the signal-to-noise ratio (SNR) evaluated by illumination and reflection angle. Secondly, we established an SNR estimated model by studying the mathematical correlations between light intensity, reflected angle, and SNR of the regions on the face. Then, we proposed a robust ROI selection method based on SNR estimated model for adaptive ROI selection. Finally, the ROI change de-noising method is proposed to remove the ROI change noise for synthesizing the imaging Photoplethysmography (iPPG) signal and calculating the heart rate. Performances of the proposed method were tested on three datasets, including our collected dataset and two public datasets. The experimental results show that on the MR-NIRP CAR dataset, the percentage of the time that the heart rate (HR) error is less than 6 bpm (PET6) value is higher than 7.3% and 4.6%, respectively, compared with the State-of-the-art (SOTA) methods. The SNR, standard deviation error (SD), mean absolute error (MAE), and root mean squared error (RMSE) metrics are also improved on the Head-rotation and PURE datasets. Therefore, our method is robust and can potentially be applied in various scenarios.
- 13 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72 (00189456) , 1-12
Robust Heart Rate Measurement by Adaptive ROI Selection for Head-Rotation Conditions
Region of interest (ROI) selection is a key process for remote heart rate measurement approach. Previous studies have shown that the selected ROI could easily be lost under head rotation circumstances and does not perform well under drastic changes in illumination. To solve the problems, this work proposed a novel adaptive ROI selection method. Firstly, we collected a dataset for modeling the signal-to-noise ratio (SNR) evaluated by illumination and reflection angle. Secondly, we established an SNR estimated model by studying the mathematical correlations between light intensity, reflected angle, and SNR of the regions on the face. Then, we proposed a robust ROI selection method based on SNR estimated model for adaptive ROI selection. Finally, the ROI change de-noising method is proposed to remove the ROI change noise for synthesizing the imaging Photoplethysmography (iPPG) signal and calculating the heart rate. Performances of the proposed method were tested on three datasets, including our collected dataset and two public datasets. The experimental results show that on the MR-NIRP CAR dataset, the percentage of the time that the heart rate (HR) error is less than 6 bpm (PET6) value is higher than 7.3% and 4.6%, respectively, compared with the State-of-the-art (SOTA) methods. The SNR, standard deviation error (SD), mean absolute error (MAE), and root mean squared error (RMSE) metrics are also improved on the Head-rotation and PURE datasets. Therefore, our method is robust and can potentially be applied in various scenarios.