Table 8 Comparison of related works with our method.

Monte Moreno¹² (2011)

410

87.71

10.32

Linear Regression / Support Vector Machine / Artificial Neural Network / Random Forest

Physiological features, features derived from PPG, and other vital signals

9-80 Mean = 37.9, SD = 13.3

No

University personnel and ambulatory medical assistance staff

J. Yadav et al.²⁷ (2017)

50

86.01

13.99

Multi Linear Regression, Artificial Neural Network

Physiological features, features derived from PPG, and other vital signals

21 - 30 Mean = 24, SD = 3

No

Collected by Author

R. Bunescu et al.⁴² (2013)

10

NA

NA

Support Vector Machine

The dynamics of meal absorption, / insulin, and glucose, along with a feature generated using ARIMA modeling

NA

No

Collected by Author

S. Ramasaha yam²¹ (2015)

55

95.38

4.6

Artificial Neural Network

Measurements of light absorption intensities

NA

FPGA implementation

NA

S. Habbu²⁸ (2019)

611

83.0

17.0

Artificial Neural Network

Features derived from PPG

4-70

No

Jahangir Medical and Research Centre, India

P. Jain et al.²⁵ (2019)

190

97.0

3.0

Deep Neural Network

PPG

17-77

ML implementation on Arduino

Collected by Author

Shantanu Sen Gupta et al.⁴³ (2021)

26

96.0

3.85

Random Forest, XGBoost

17 features derived from PPG

25-80 Mean = 30.31, SD=2.38

No

Collected by Author

J. Chu et al.²⁶ (2021)

2538

60.6

37.4

1d CNN with micro and macro training

Raw PPG

38 - 80 Mean = 63.15, SD = 9.67

No

Institutional Review Board of Academia Sinica, Taiwan

Z. Nie et al.⁴⁴

8

89.6

10.4

Machine learning

IPPG, NIR, Feature extraction, RFR

20-35

No

Collected by Author

Shisen Chen et al.⁴⁵ (2024)

260

87.39

12.11

Deep Neural Network

PPG kinetic features, PPG Derivatives

16-82 Mean= 43, SD= 13.8

No

Collected by Author

Our Work

6388(train + test) + 67(test)

72.6

25.9

Deep Neural Network (CNN)

Raw PPG

0.3-94 Mean= 58.8, SD= 15.1

Using STM32

VitalDB (train and test) + MUST (test)