This study presents a non-invasive approach for detecting blood glucose levels using smartphone image processing. Venous blood samples were taken from 30 participants and their hand skin images were captured using a smartphone camera. Images were preprocessed and GLCM texture features were extracted. An artificial neural network was trained to correlate glucose levels with texture values. A mobile app called GULAABLE was developed to analyze images and output glucose levels as low or high. When tested on 10 patients, the app achieved 80% accuracy compared to invasive measurements. The study demonstrates the potential for non-invasive blood glucose monitoring using smartphone imaging and machine learning.

1. Presented by
<Usman Umar>
International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
<Usman Umar, Risnawaty Alyah, Mustapa>
A Non-invasive Approach to Detection
Blood Glucose Levels with Image
Processing Using Smartphone
<319>

2. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Introduction
1/17
Blood and glucose is a very important components in human body tissues
Glucose is a carbohydrate element that produces an energy source for all body
cell tissues, accelerates metabolism and functions as the main fuel for the
brain, and controls body temperature
Uncontrolled blood glucose conditions can cause blood vessel disease
Excessive glucose levels over a long period of time can lead to diabetes, which
can be complicated by other diseases such as nerve damage, vision loss,
kidney damage, kidney disorders, and an increased risk of cardiovascular
disease
The device for measuring total glucose available in general clinical laboratories
is the one with invasive techniques.
Invasive technique procedures require blood samples collection which
poses a risk of bruising and inflammation

3. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Related Works
2/17
Research using measurement methods with non-invasive techniques such as
V. P. Rachim and
W. Y. Chung
(2018)
• Noninvasive blood glucose monitor via
multi-sensor fusion and its clinical
evaluation.
S. Ghosal, A.
Kumar, V.
Udutalapally,
and D. Das
(2019)
• Glucam: Smartphone based blood glucose
monitoring and diabetic sensing.
H. Zhang, Z.
Chen, J. Dai, W.
Zhang, Y. Jiang,
and A. Zhou
(2020)
• A low-cost mobile platform for whole blood
glucose monitoring using colorimetric method
F. Rui, G.
Zhanxiao, L. Ang,
C. Yao, W.
Chenyang, and
Z. Ning (2021)
• A low-cost mobile platform for whole blood
glucose monitoring using colorimetric
method

4. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Proposed Approach
3/17
The proposed study describes a blood glucose detection system based on hand
skin image processing under Artificial Neural Networks (ANN). which is
implemented on a smartphone with QS android through the GULAABLE
application.
Invasive Method
Non-Invasive Method

5. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Experimental Details
4/17
ARCHITECTURAL DESIGN
The initial step in this study was
venous blood sampling for invasive
glucose level measurement from 30
participants with an age range of 20 -
60 years
The second step is taking the image of
the skin of the hand with 4 hand
positions, the image of the skin of the
hand is taken with a 13 megapixel
smartphone camera with an object
distance of 10 cm.

6. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Experimental Details
5/17
Meassurement glucose levels with invasive method
NO Participant GLUCOSE NO Participant GLUCOSE
mg/dl mg/dl
1 Sample 1 84 11 Sample 11 146
2 Sample2 184 12 Sample 12 90
3 Sample 3 73 13 Sample 13 93
4 Sample 4 103 14 Sample 14 224
5 Sample 5 132 15 Sample 15 138
6 Sample 6 90 16 Sample 16 75
7 Sample 7 111 17 Sample 17 101
8 Sample 8 119 18 Sample 18 79
9 Sample 9 81 19 Sample 19 218
10 Sample 10 75 20 Sample 20 90

7. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Experimental Details
6/17
Pre-processing Image
The next step is image pre-processing to reduce noise and unnecessary
information from the image and reduce variations that arise when the image
is captured, then cropping the image with the "im-crop algorithm" and the
data is stored as an image database.

8. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Experimental Details
7/17
Gray Level Co-occurrence matrix (GLCM) Texture Extraction

9. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Experimental Details
8/17
Gray Level Co-occurrence matrix (GLCM) Texture Extraction
No Participant
GLCM Feature Extraction Value
No Participant
GLCM Feature Extraction Value
Contrast Correlation Energy Homogeneity Contrast Correlation Energy Homogeneity
1 Sampel 1 0.098555 0.88196 0.34667 0.95075 11 Sampel 11 0.07134 0.90859 0.37779 0.96433
2 Sampel 2 0.086571 0.89442 0.38517 0.95673 12 Sampel 12 0.084665 0.87291 0.39293 0.95767
3 Sampel 3 0.086148 0.8382 0.46703 0.95693 13 Sampel 13 0.067097 0.9357 0.36943 0.96646
4 Sampel 4 0.067255 0.95551 0.28309 0.96637 14 Sampel 14 0.10713 0.85561 0.37686 0.94673
5 Sampel 5 0.03198 0.77889 0.83127 0.98401 15 Sampel 15 0.06168 0.92622 0.38261 0.96916
6 Sampel 6 0.073775 0.89455 0.41545 0.96311 16 Sampel 16 0.11783 0.84175 0.38102 0.94112
7 Sampel 7 0.083086 0.87618 0.41396 0.95849 17 Sampel 17 0.082532 0.89421 0.38552 0.95874
8 Sampel 8 0.093749 0.95264 0.23283 0.95338 18 Sampel 18 0.045529 0.90466 0.51469 0.97724
9 Sampel 9 0.09077 0.92559 0.28675 0.95462 19 Sampel 19 0.087072 0.92875 0.29694 0.95651
10 Sampel 10 0.078451 0.91136 0.34591 0.96078 20 Sampel 20 0.037927 0.97613 0.30296 0.98104

10. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Experimental Details
9/17
ARTIFICIAL NEURALNETWORK (ANN)
Training data using ANN with
backpropogation algorithm. by
linearly correlating between
invasive glucose levels and
GLCM texture values at each
angle at 4 image positions that
have been cropped with a size of
1000 pixels from 20 participants.
Training data

11. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Experimental Details
10/17
GULAABLE app
The final stage of developing GULAABLE on an android smartphone to analyze
blood glucose calcification.
The data base training data with ANN is uploaded to the android operating
system for the GULAABLE application.
Then the GULAABLE app is used to test the data with glucose calcification
output LOW or HIGH.

12. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Results and Analysis
11/17
The results of the plot, show a
regression coefficient (R) of
0.91397.
R-value shows the relationship
between GCLM values and
glucose levels. very strong for
predicting total glucose because it
is close to 1

13. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Results and Analysis
12/17
Evaluation Model
Data testing program using the GULAABLE application on a smartphone
The first step is to install
and open the GULAABLE
application on a
smartphone
The second step of the
menu display gives the
option to add the image of
the skin of the hand with
the choice of taking
pictures from files
Then the third step after the
skin image has been
obtained, cropping is carried
out according to the required
size, and the cropping results
will be displayed for analysis

14. International Conference on Advanced Computing and Intelligent Engineering (ICACIE 2016)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Results and Analysis
13/17
Evaluation Model
Data testing program using the GULAABLE application on a smartphone
The next stage is the analysis process, this process
takes about 10 minutes, depending on the size of
the cropping image. The results of the analysis are
based on the displayed LOW or HIGH blood
glucose conditions
Finally step is finalization by
editing or inputting the
patient's identity and saving it.

15. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Results and Analysis
14/17
Evaluation Model
No
Name, Gender and Age
(year)
CONDITION
Glucose
Levels
Invasive
Result
GULAABLE
1
2
3
4
5
6
7
8
9
10
Rohana, (P.50)
Mardiana .( P. 23)
Eda, (P.44)
Hadayati, (P. 59)
Kamisa (P.43)
Ahmad (L.50)
Sri Rahayu. P.50
Ramli D.S , L.60
Kurnia, P.41
Hakim Rewa L.39
Normal
Normal
Fasting
Fasting
Fasting
Normal
Fasting
After eat
Normal
Normal
304
218
81
75
224
194
160
211
138
266
HIGH
HIGH
LOW
LOW
HIGH
HIGH
HIGH
HIGH
HIGH
HIHG
The results of the identification of the gulaable app are compared with the results of
invasive glucose measurements

16. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Results and Analysis
15/17
Evaluation Model
This study develops the GULAABLE application to identify blood glucose levels which
identify blood glucose levels through image processing of hand skin images
The results of this study show results with good accuracy, proving that non-invasive
monitoring of blood glucose levels allows detection through the skin of the hands.
From the table of test results on 10 patients. showed 80% accuracy, misidentification
in 2 patients out of 10 patients

17. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
Conclusions
16/17
The proposed research has developed an innovative intelligent control application
to detect glucose levels. In this study, blood glucose levels can be detected by
image processing that is applied to a smartphone.
The correlation between invasive glucose levels and GLCM extraction values in
artificial neural network (ANN) regression plots using the backpropagation method
showed a very strong relationship with the R-value; 0.91 is close to 1.
The development of the GULAABLE application on a smartphone to detect glucose
with a non-invasive technique, where the results of LOW or HIGH calcification of
participants' glucose conditions are compared with the results of laboratory tests,
from the analytical data describing an acceptable accuracy to be applied to the
community

18. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
References
17/17
[1] W. Zheng et al., “Highly-sensitive and reflective glucose sensor based on optical fiber surface plasmon
resonance,” Microchem. J., vol. 157, no. February, p. 105010, 2020, doi:
10.1016/j.microc.2020.105010.
[2] X. Dong et al., “Influence of blood glucose level on the prognosis of patients with diabetes mellitus
complicated with ischemic stroke,” no. 65, pp. 1–7, 2018, doi: 10.4103/1735-1995.223951.
[3] K. S. P and S. Am, “A study on the glycemic , lipid and blood pressure control among the type 2
diabetes patients of north Kerala , India,” Indian Heart J., vol. 70, no. 4, pp. 482–485, 2018, doi:
10.1016/j.ihj.2017.10.007.
[4] A. Kerimi, H. Nyambe, S. Alison, P. Ebun, O. Julia, and S. G. Yala, “Nutritional implications of olives
and sugar : attenuation of post- prandial glucose spikes in healthy volunteers by inhibition of sucrose
hydrolysis and glucose transport by oleuropein,” Eur. J. Nutr., vol. 58, no. 3, pp. 1315–1330, 2019, doi:
10.1007/s00394-018-1662-9.
[5] C. Beehan-quirk et al., “Investigating the effects of fatigue on blood glucose levels – implications for
diabetes,” Transl. Metab. Syndr. Res., 2020, doi: 10.1016/j.tmsr.2020.03.001.
[6] K. Ogurtsova et al., “IDF Diabetes Atlas: Global estimates for the prevalence of diabetes for 2015 and
2040,” Diabetes Res. Clin. Pract., vol. 128, pp. 40–50, 2017, doi: 10.1016/j.diabres.2017.03.024.
[7] N. H. Cho et al., “IDF Diabetes Atlas : Global estimates of diabetes prevalence for 2017 and projections
for 2045,” Diabetes Res. Clin. Pract., vol. 138, pp. 271–281, 2018, doi: 10.1016/j.diabres.2018.02.023.
[8] Y. Sun, Y. Song, C. Liu, and J. Geng, “Saudi Journal of Biological Sciences Correlation between the
glucose level and the development of acute pancreatitis,” Saudi J. Biol. Sci., vol. 26, no. 2, pp. 427–
430, 2019, doi: 10.1016/j.sjbs.2018.11.012.

19. International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)
Introduction
Related Works
Proposed Approach
Experimental Details
Results and Analysis
Conclusions
References
18/17
[9] S. R. Chinnadayyala, J. Park, A. T. Satti, D. Kim, and S. Cho, “Minimally invasive and
continuous glucose monitoring sensor based on non-enzymatic porous platinum black-coated
gold microneedles,” Electrochim. Acta, vol. 369, p. 137691, Feb. 2021, doi:
10.1016/j.electacta.2020.137691.
[10] N. Duc, T. Van Nguyen, A. Duc, and H. Vinh, “ORIGINAL
ARTICLE A label-free colorimetric sensor based on silver
nanoparticles directed to hydrogen peroxide and glucose,” Arab. J
. Chem., vol. 11, no. 7, pp. 1134–1143, 2018, doi
[11] F. Rui, G. Zhanxiao, L. Ang, C. Yao, W. Chenyang, and Z. Ning, “Sensors and Actuators : B .
Chemical Noninvasive blood glucose monitor via multi-sensor fusion and its clinical
evaluation,” Sensors Actuators B. Chem., vol. 332, no. September 2020, p. 129445, 2021, doi:
10.1016/j.snb.2021.129445.
[12] V. P. Rachim and W. Y. Chung, “Wearable-band type visible-near infrared optical biosensor for
non-invasive blood glucose monitoring,” Sensors Actuators, B Chem., vol. 286, no. October
2018, pp. 173–180, 2019, doi: 10.1016/j.snb.2019.01.121.
[13] H. Zhang, Z. Chen, J. Dai, W. Zhang, Y. Jiang, and A. Zhou, “A low-cost mobile platform for
whole blood glucose monitoring using colorimetric method,” Microchem. J., vol. 162, no.
October 2020, p. 105814, 2021, doi: 10.1016/j.microc.2020.105814.
[14] S. Ghosal, A. Kumar, V. Udutalapally, and D. Das, “glucam: Smartphone based blood glucose
monitoring and diabetic sensing,” IEEE Sens. J., vol. 21, no. 21, pp. 24869–24878, 2021.

20. ?
Q and A?
International Conference on Electronics, Biomedical Engineering, and Health Informatics (iCEBEHI 2022)