=Paper=
{{Paper
|id=Vol-3789/Paper8
|storemode=property
|title=AI to minimise human errors in the detection of hematologicaldiseases
|pdfUrl=https://ceur-ws.org/Vol-3789/Paper8.pdf
|volume=Vol-3789
|authors=Bana Fridath BIO NIGAN,Alban Gildas ZOHOUN,Ahmed Dooguy KORA
|dblpUrl=https://dblp.org/rec/conf/cita2/NiganZK24
}}
==AI to minimise human errors in the detection of hematologicaldiseases==
https://ceur-ws.org/Vol-3789/Paper8.pdf
AI to minimise human errors in the detection of hematological
diseases⋆
Bana Fridath BIO NIGAN1,∗,† , Alban Gildas ZOHOUN2,† and Ahmed Dooguy KORA1,†
1
Laboratory E-Inov, Ecole Supérieure Multinationale des Télécommunications, Dakar – SENEGAL
2
Laboratoire d’hématologie, Faculté des Sciences de Santé – CNHU-HKM, Cotonou – BENIN
Abstract
Hematology is a branch of medicine that relies on accurate diagnosis and appropriate treatment of blood-related diseases. However,
human errors, whether due to technician fatigue, inattention or technical limitations, can have serious consequences for patients.
Artificial intelligence offers a solution to these problems. By integrating advanced machine learning and deep learning algorithms, AI
offers innovative solutions for reducing diagnostic and treatment errors. With its ability to analyse big data with high accuracy, AI
promises to transform hematology practice, ensuring safer and more effective care for patients. This article reviews the different AI
techniques used in the recognition of blood cells and the detection of related diseases, while highlighting its benefits in minimising
human errors in diagnosis.
Keywords
AI, errors, hematology, minimise.
1. Introduction – Mononuclear cells (lymphocytes, monocytes).
• Platelets (PLT), which are anucleate fragments and
Artificial intelligence (AI) is defined as a process of imitating
occur at a rate of 150,000 to 450,000/mm3.
human intelligence, based on the creation and application
of algorithms executed in a dynamic computer environment Each type of cell has its own distinguishing features,
[1]. It is therefore the ability of a machine to imitate human whether in terms of shape, colour or even size.
behavior (analysis, interpretation, decision-making based In most of our hematology laboratories in Africa and in
on an image) via algorithms and to make predictions based Benin in particular (CNHU-HKM), our technicians carry
on data already acquired [2]. AI is faster and makes fewer out this recognition work manually using a microscope. In
errors than humans when it comes to achieving results. Its addition to the long wait by patients before obtaining re-
importance is growing by the day. It has developed and sults, these results are sometimes exposed to human errors
is now present in almost every sector of human activity: of inattention. This is justified by the large number of blood
transport, agriculture, commerce, medicine, .... slides to be analysed by these technicians.
The use of AI in medicine began in the 20th century By using advanced algorithms and massive data process-
in developed countries for the rapid management of ing capabilities, AI offers unprecedented opportunities to
patients and the accurate diagnosis of certain diseases improve the diagnosis, treatment and management of hema-
[3]. Today, AI is commonly used in the detection of rare tological diseases and sometimes reduce the cost of analysis.
genetic diseases (Cornelia de Lange syndrome, Angelman This is the case of an AI model used by some authors to help
syndrome, etc.) [4], heart disease and cancer [5], blood practitioners identifying different hematological diseases
diseases, …. It is transforming many areas of medicine, with inexpensive hemogram tests. This binary and multi-
including hematology. Hematology is a medical speciality class classification model achieved up to 96% accuracy [8].
that studies blood, the hematopoietic organs (bone marrow, This article explores the different applications of AI in hema-
lymph nodes and spleen being the main ones) and their tology, highlighting the potential benefits of this revolution-
diseases [6]. ary technology.
Blood includes blood cells in plasma which are made in We first present the litterature review on recognition and
the red bone marrow from a stem cell. By dividing and classification of blood cells and automatic detection of blood
differentiating, this cell gives rise to one of three categories diseases. Then, we present the CNN model designed for the
of blood cells [7]: CNHU- HKM hematology laboratory and performances ob-
tained. Finally, we discuss the benefits of AI in minimising
human error in hematology diagnosis.
• Red blood cells (RBC), also known as erythrocytes,
which are anucleate cells and are the most numerous,
around 5 million/mm3. 2. Litterature Review
• White blood cells (WBC) or leukocytes, around
8000/mm3: 2.1. Recognition and Classification of Blood
– Polynuclear cells or granulocytes (neu- Cells
trophils, basophils, eosinophils);
2.1.1. White blood cells (WBC)
Cotonou’24: Conférence Internationale des Technologies de l’Information
et de la Communication de l’ANSALB, June 27–28, 2024, Cotonou, BENIN
⋆
For cell recognition, authors used different ML/DL methods
You can use this document as the template for preparing your publica- for cell segmentation, classification, and counting.
tion. We recommend using the latest version of the ceurart style.
∗ S Khan et al. (2021) used both traditional learning meth-
Corresponding author.
†
These authors contributed equally.
ods (manual extraction of features + classification cells by
Envelope-Open fridabionigan@gmail.com (B. F. B. NIGAN) ANN) and DL-based methods (characteristic extraction +
Orcid 0000-0001-9950-8821 (B. F. B. NIGAN) classification cells by CNN) to classify WBCs. This study
© 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License
Attribution 4.0 International (CC BY 4.0). reveals that they achieved the same performance in all 02
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
cases [9]. RB Hegde et al. (2019) also compared traditional tocytes, echinocytes, lacrimal cells and macrocytes) with
and DL methods and arrived at the same results with 99% an accuracy of 91.667% over a period of 0.81432 seconds for
of accuracy for WBCs classification [10]. different blood samples [25].
AS Ashour et al. (2021) used a neural network associated Namata et al (2021) proposed an image processing method
with the SVM algorithm on a database combination and using a convolution neural network for classification of
achieved a segmentation performance of 94.9counting ac- RBCs. The algorithm used extracts features from segmented
curacy for both cells is 97.4Ruberto et al. (2015) created a images and classifies in nine categories with an overall ac-
multi-classifier system for WBC segmentation using ANN curacy of 98.5% [26].
+ Nearest Neighbor and SVM algorithms and obtained a
segmentation accuracy of 99% [12]. 2.2. Automatic detection of blood cells
S Manik et al. (2016) used an ANN-based system with MAT-
LAB capabilities to detect and classify WBC. The accuracy
diseases
of the entire system is 98.9% with 100for eosinophils and 2.2.1. White blood cells
neutrophils, 96.7% for lymphocytes [13]. Using the Faster
R-CNN (Fast R-CNN + Region Proposal Network (RPN)) The excessive presence of certain immature cells in the pe-
method on the BCCD dataset to recognize and classify dif- ripheral blood reveals that the patient has a disease, the case
ferent blood cells, S Raina et al. (2020) obtained the following of blasts for leukemia [27].
results: RBC 55.83%, PLT 68.36% and WBC 92.10% [14]. M Jiang et al. (2018) developed a WBCNet model to fully
J Basnet et al. (2020), using a Deep CNN-based method, extract WBC characteristics by combining a batch normali-
improved classification accuracy from 96.1% to 98.92% and sation algorithm, residual convolution architecture and the
reduced processing time from 0.354 to 0.216 s [15]. MJ enhanced activation function to diagnose leukemia and re-
Macawile et al. (2018) created a CNN- based system to clas- duce the misdiagnosis rate. This model obtained an accuracy
sify and count WBCs using the HSV (Hue Saturation Value) of 77.65% and 98.65% for Top-1 and Top-5 respectively in
saturation component on the ALL-IDB database. They com- training and 83% in testing for Top-1 [28]. To improve these
pared several models (Alexnet, GoogleNet, and ResNet-101). results, Sheikh IM Chachoo, MA. (2020) used an advanced
AlexNet appears to be the best, with a sensitivity of 89.18%, ML-based method to segment the GBs. This segmentation
a specificity of 97.85% and an accuracy of 96.63% [16]. is based on grey level and consists of eliminating the other
With a model based on the Deep RN and using the character- cells and the cytoplasm of the WBCs and extracting only
istics of the convolutional layers of the AlexNet architecture, their nuclei. It achieved a nucleus extraction accuracy of
A Khan et al. (2021) were able to identify the different types 91%. This method is only applicable to WBCs [29].
of WBCs with a training accuracy of 99.99% and the test of An automatic CNN system has been designed by Anwar
99.12%. This model is named MLANet-FS-ELM [17]. S, Alam A (2020) for the detection of acute lymphoblastic
A Malkawi et al. (2020) set up a CNN-based hybrid system leukemia (ALL) without preprocessing or segmentation. It
to extract WBCs characteristics and classify them. They has achieved 99.5% accuracy [30]. Boldú L et al. (2021)
evaluated the performance of 3 classifiers (SVM, k-NN, RF) used a LD model to firstly recognise lymphocytes, mono-
on the LISC WBC and the RF performed better with a test cytes, blasts and activated lymphocytes and then classify
accuracy of 98.7% [18]. A. Şengür et al. (2019) used a system blasts found. Authors ran several architectures (VGG16,
based on image processing and ML in particular Deep CNN ResNet101, DenseNet121, SENet154, ALNet (02 CNNs in
to classify WBCs according to their shape and other deep series)). The ALNet model performed better: Myeloid
characteristics. This system achieved an accuracy of 80% leukemia (accuracy 93.7%, specificity 92.3%, sensitivity 100%)
in relation to the shape and 82.9% in relation to the deep and Lymphoid leukemia (accuracy & specificity 100sensitiv-
characteristics; combining these 02 parameters, the overall ity 89%) [31].
accuracy is 85.7 An ML model based on digital image processing techniques
and the RF classifier has enabled Mohamed H et al. (2018)
to diagnose WBC-related diseases. The model achieved an
2.1.2. Red blood cells
accuracy of 94.3% [32]. Similarly, another study conducted
Differents ML/DL techniques have been used for automatic by Sheng B et al. (2020), used the Faster R-CNN method
cell recognition [20]. combined with the VGG16 technique to classify WBCs and
Maity et al. (2012) employed an efficient supervised- detect the presence of lymphoma in the blood. It obtained a
decision-tree C4.5 to classify RBCs into six sub-classes in- lymphoma detection rate of > 96% [33].
cluding sickle-cells with 98.2% precision and 99.6% speci- Agrawal R et al. (2019) developed a CNN model for diag-
ficity [21]. They also proposed another method which em- nosing all types of cancer. Its operation is based on image
phasizes the extraction of crucial shape-based features for processing techniques. The system is 97.3% accurate [34]. A
RBC classification into nine classes including healthy cells decision support system based on ANNs was used by Negm
in 2017. This method achieved 99.71% specificity and 97.81% AS (2018) to identify blasts. Several classifiers (k-Means,
accuracy [22]. LBG, KPE) were evaluated and k-Means performed better
Acharya and Kumar (2017) employed a technique capable to with an accuracy of 99.74% and a sensitivity of 100% [35].
classify RBCs into 11 sub-classes including sickle-cells with
98% precision [23]. Using the Hough Circular Transform 2.2.2. Red blood cells
(HCT) method, Mazalan SM et al. (2013) were able to count
the total number of RBCs in a peripheral blood smear im- The literature is full of studies on the detection of diseases
age. Results showed that from ten sample peripheral blood caused by RBCs.
smear images, accuracy was 91.87% [24]. Using the same Normal and abnormal cells were classified into four classes:
method, Chadha GK et al. (2020) were able to count and sickle cells, dacrocytes, ovalocytes and erythrocytes by
classify RBCs according to four types of abnormality (ellip- Sharma V (2016) using the KNN classifier and Watershed
segmentation technique with an accuracy of 80.6% [36]. Xu
M et al. (2017) focused on the RBC shape detection using
different techniques. A Deep CNN was used to find their
region of interest (ROI) using an automatic seed generation
technique and a mask based on patch normalization to ob-
tain images of uniform size. This method is not widely used
because it requires a very large database [37].
Sobel’s edge detection algorithm is used by Mohamad A et
al. (2017) for detecting RBC shape with blob measurement.
This inexpensive technique is beneficial for people living
in remote areas and achieveded 95% accuracy but only for
2D images [38]. Zhang M et al. (2020) adopted a semantic
segmentation framework based on deep learning to solve
the GR classification task. The performance obtained was
Figure 2: Accuracy curve 16*16
97% for the dU-Net model and 94.7% for the classical U-Net
model [39].
A transfer learning technique that automatically extracts The loss curve evaluates how well our algorithm models
features and is specific to small databases has also been the dataset. The lower the loss, the better is. Figure3 and 4
proposed by Alzubaidi L et al. (2020). Thanks to data aug- show performance obtained in the automatic classification
mentation, it achieved 99.98% [40]. Chy T et al. (2019) used of blood cells for 32*32 images and 16*16 images respectively.
different techniques (fuzzy C mean clustering algorithm, We notice that while the training curve tends towards zero
KNN, SVM and ELM) to automatically detect sickle cell dis- (0), the validation curve is a bit high [45].
ease. ELM classifier performed better, with an accuracy of
95.45% [41]. AlexNet was also used by Aliyu H. et al. (2020)
to classify red blood cells in sickle cell anemia. The accuracy
obtained was 95.92% [42]. Another technique using a smart-
phone microscope has also been used on blood smears for
the same purpose by De Haan K. et al. (2020). It comprises
two distinct and complementary deep neural networks and
achieved an accuracy of around 98% [43].
3. Performances of the proposed
model for the CNHU-HKM
hematology laboratory
The model designed for CNHU-HKM hematology laboratory Figure 3: Loss curve 32*32
is a CNN architecture of 12-layers. It allows to recognize
blood cells (WBC, RBC) and detect blood cells diseases like
sickle cell anemia in CNHU-HKM hematology laboratory
[44]. These layers are optimized for the maximum positive
prediction rate. Figure1 and 2 show performance obtained
in automatic classification of blood cells for 32*32 images
and 16*16 images respectively with accuracy of 98.78% for
training and 88.11% for test and 86.59% for cross validation
for 32*32 images, and 90.11% for test and 88.53% for cross
validation with 16*16 images [44].
Figure 4: Loss curve 16*16
Figure5 and 6 show performance obtained in automatic
detection of sickle cells disease, elliptocytosis and schizocy-
tosis for 32*32 images and 16*16 images respectively. After
running the model, we obtained 100% for training and 82%
for validation with 32*32 images, and 86% for validation
with 16*16 images [45].
Figure 1: Accuracy curve 32*32
in terms of regulation, ethics and acceptance by healthcare
professionals, the potential benefits of AI are immense. By
improving diagnostic accuracy, optimizing treatments, and
facilitating data management, AI has the potential to reduce
human error and improve patient outcomes. The future of
hematology, enriched by AI, looks promising and will bring
significant innovations.
6. Acknowledgments
We thank Professor Issiako Bio Nigan for his recommenda-
tions. We also thank the CNHU-HKM hematology labora-
Figure 5: Accuracy curve 32*32
tory of Benin for giving us access to data.
References
[1] Parlement européen. Intelligence artificielle:
définition et utilisations. [Online]. 2023. Available:
https://www.europarl.europa.eu/news/fr/head-
lines/society/202 00827STO85804/intelligence-
artificielle-definition-et- utilisation.
[2] UNESCO, Site officiel de la Chaire UNESCO. Qu’est-
ce- que l’intelligence artificielle ? [Online]. 2023.
Available: https://chaireunesco.org/Intelligence_ ar-
tificielle.
Figure 6: Accuracy curve 16*16 [3] Callier P, Sandel O. De l’intelligence artificielle
à son application en médecine. Actualités phar-
maceutiques. [Online]. 2021. P. 18-20. Avail-
4. Discussion able: https://www.sciencedirect.com/science/arti-
cle/abs/pii/S05153 70021004055.
AI has remarkable potential to reduce human error in hema-
[4] Labbe P. Une intelligence artificielle qui détecte
tology. It is beneficial in several ways:
les maladies génétiques rares. [Online]. Avail-
• Data analysis: ML algorithms can analyse big med- able: https://technplay.com/ia-maladies-genetiques-
ical data, including blood test results, to detect rares/.
anomalies or patterns that humans might miss. [5] Hughes P. Sama game – l’arène des nouvelles sur
This can help diagnose hematological diseases more les jeux Français. [Online]. 2023. Available:
quickly and accurately. https://samagame.com/fr/news/.
• Clinical decision support: AI systems can provide [6] Marion Spée. Hématologie. [Online]. 2016. Avail-
recommendations based on clinical data and best able: https://www.passeportsante.net/fr/specialites-
practice, helping doctors to make more informed de- medicales/Fiche.aspx?doc=hematologie.
cisions. For example, in hematology, AI can suggest [7] S. Web. « Le sang (hématies, plaquettes...) circule
optimal treatments for diseases such as leukemia or dans les vaisseaux sanguins ». [Online]. Available:
anemia. https://www.toutsurlatransfusion.com/transfusion-
• corresponding author mark : sanguine/medecine-transfusionnelle/composition-
\cormark[] Automation of repetitive du-sang.php.
tasks: AI can automate administrative and clinical [8] Sergio Diaz-del-Pino, Roberto Trelles-Martinez, F.A.
tasks, reducing the risk of human error. For example, González-Fernández, Nicolas Guil. Artificial in-
the transcription of test results or the management telligence to assist specialists in the detection of
of medical records can be carried out by AI systems, haematological diseases. 2405-8440/© 2023 Pub-
which relieves healthcare professionals. lished by Elsevier Ltd. https://doi.org/10.1016/j.he-
• Continuous monitoring: AI systems can continu- liyon.2023.e15940.
ously monitor patients and alert doctors to signif- [9] S. Khan, M. Sajjad, T. Hussain, A. Ullah, et A. S.
icant changes in health parameters. This is partic- Imran, « A Review on Traditional Machine Learn-
ularly useful for patients with chronic diseases or ing and Deep Learning Models for WBCs Classi-
who require constant monitoring, such as certain fication in Blood Smear Images », IEEE Access,
hematological diseases. vol. 9, p. 10657‑10673, 2021, doi: 10.1109/AC-
CESS.2020.3048172.
[10] R. B. Hegde, K. Prasad, H. Hebbar, et B. M. K. Singh,
5. Conclusion « Comparison of traditional image processing and
deep learning approaches for classification of white
In conclusion, AI integration in hematology represents a
blood cells in peripheral blood smear images », Bio-
major advance that promises to transform clinical and re-
cybernetics and Biomedical Engineering, vol. 39, no
search practices. Although challenges remain, particularly
2, p. 382‑392, 2019, doi: 10.1016/j.bbe.2019.01.005.
[11] A. S. Ashour, M. A. Wahba, et R. Ghannam, « A [21] M. Maity, P. Sarkar and C. Chakraborty. ”Computer-
Cascaded Classification-Segmentation Reversible assisted approach to anemic erythrocyte classifica-
System for Computer-Aided Detection and Cells tion using blood pathological information,” in Proc.
Counting in Microscopic Peripheral Blood Smear 3rd Int. Conf. Emerg. Appl. Inform. Technol., 2012,
Basophils and Eosinophils Images », IEEE Access, pp. 116-121.
vol. 9, p. 78883‑78901, 2021, doi: 10.1109/AC- [22] M. Maity, T. Mungle, D. Dhane, A.K. Maiti and C.
CESS.2021.3083703. Chakraborty. ”An ensemble rule learning approach
[12] C. Di Ruberto, A. Loddo, et L. Putzu, « A Multiple for automated morphological classification of ery-
Classifier Learning by Sampling System for White throcytes,” J. Med. Syst., vol. 41, pp. 41-56, 2017.
Blood Cells Segmentation », in Computer Analysis [23] V. Acharya and P. Kumar. ”Identification and red
of Images and Patterns, G. Azzopardi et N. Petkov, blood cell classification using computer aided system
Éd., in Lecture Notes in Computer Science, vol. 9257. to diagnose blood disorders,” in Proc. Int. Conf. adv.
Cham: Springer International Publishing, 2015, p. Comput., Inform., 2017, pp. 2098-2104.
415‑425. doi: 10.1007/978-3-319-23117-4-36. [24] Mazalan SM, Mahmood NH, Razak MAA, dir. Au-
[13] S. Manik, L. M. Saini, et N. Vadera, « Counting and tomated red blood cells counting in peripheral
classification of white blood cell using Artificial Neu- blood smear image using circular hough transform.
ral Network (ANN) », 1st International Conference 1st International Conference on Artificial Intelli-
on Power Electronics, Intelligent Control and En- gence, Modelling and Simulation; 3-5; Kota Kinabalu,
ergy Systems (ICPEICES), Delhi, India: IEEE, 2016, Malaisie. IEEE; 2013. [DOI 10.1109/AIMS.2013.59].
p. 1‑5. doi: 10.1109/ICPEICES.2016.785364. [25] Chadha GK, SrivastavaA, Singh A, Gupta R, Singla D.
[14] S. Raina, A. Khandelwal, S. Gupta, et A. Leekha, « An automated method for counting red blood cells
Blood Cells Detection Using Faster-RCNN », Interna- using image processing. Procedia Computer Science.
tional Conference on Computing, Power and Com- 2020; 167: 769‑78. [DOI 10.1016/j.procs.2020.03.408].
munication Technologies (GUCON), Greater Noida, [26] Mamata AP, Ninad DM. Red Blood Cell Classifcation
India: IEEE, 2020, p. 217‑222. doi: 10.1109/GU- Using Image Processing and CNN. SN Computer
CON48875.2020.9231134. Science. 2021; 2:70. https://doi.org/10.1007/s42979-
[15] J. Basnet, A. Alsadoon, P. W. C. Prasad, S. A. Aloussi, 021-00458-2.
et O. H. Alsadoon, « A Novel Solution of Using [27] SM, Santé Magazine. Leucémie : symptômes,
Deep Learning for White Blood Cells Classification: diagnostic, traitements [Online]. 2021. Avail-
Enhanced Loss Function with Regularization and able: https://www.santemagazine.fr/sante/fiche-
Weighted Loss (ELFRWL) », Neural Process Lett, vol. maladie/leucemie-177379.
52, no 2, p. 1517‑1553, 2020, doi: 10.1007/s11063-020- [28] M. Jiang, L. Cheng, F. Qin, L. Du, et M. Zhang, «
10321-9. White Blood Cells Classification with Deep Convo-
[16] M. J. Macawile, V. V. Quinones, A. Ballado, J. D. lutional Neural Networks », Int. J. Patt. Recogn.
Cruz, et M. V. Caya, « White blood cell classifica- Artif. Intell., vol. 32, no 09, p. 1857006, 2018, doi:
tion and counting using convolutional neural net- 10.1142/S0218001418570069.
work », 3rd International Conference on Control [29] Sheikh IM, Chachoo, MA. Advanced machine learn-
and Robotics Engineering (ICCRE), Nagoya: IEEE, ing for leukaemia detection based on white blood
2018, p. 259‑263. doi: 10.1109/ICCRE.2018.8376476. cell segmentation. In: Badica C, Liatsis P, Kharb L,
[17] A. Khan, A. Eker, A. Chefranov, et H. Demirel, « Chahal D. Information, Communication and Com-
White blood cell type identification using multi- puting Technology. Singapore: Springer Singapore;
layer convolutional features with an extreme- 2020. p.195-207. [DOI 10.1007/978-981-15-9671-1-
learning machine », Biomedical Signal Process- 17].
ing and Control, vol. 69, p. 102932, 2021, doi: [30] Anwar S, Alam A. A convolutional neural net-
10.1016/j.bspc.2021.102932. work–based learning approach to acute lymphoblas-
[18] A. Malkawi, R. Al-Assi, T. Salameh, B. Sheyab, tic leukaemia detection with automated feature ex-
H. Alquran, et A. M. Alqudah, « White Blood traction. Med Biol Eng Comput. 2020; 58(12):
Cells Classification Using Convolutional Neural Net- 3113‑21. [DOI 10.1007/s11517-020-02282-x].
work Hybrid System », 5th Middle East and Africa [31] Boldú L, Merino A, Acevedo A, Molina A, Rodellar J.
Conference on Biomedical Engineering (MECBME), A deep learning model (ALNet) for the diagnosis of
Amman, Jordan: IEEE, 2020, p. 1‑5. doi: acute leukaemia lineage using peripheral blood cell
10.1109/MECBME47393.2020.9265154. images. Computer Methods Programs Biomed. 2021;
[19] A. Sengur, Y. Akbulut, U. Budak, et Z. Com- 202: 105999. [DOI 10.1016/j.cmpb.2021.105999].
ert, « White Blood Cell Classification Based on [32] Mohamed H, Oma R, Saeed N, Essam A, Ayman N,
Shape and Deep Features », International Artifi- Moyiy A et al, dir. Automated detection of white
cial Intelligence and Data Processing Symposium blood cells cancer diseases. First International Work-
(IDAP), Malatya, Turkey: IEEE, 2019, p. 1‑4. doi: shop on Deep and Representation Learning (IWDRL)
10.1109/IDAP.2019.8875945. ; Cairo, Égypte. Cairo: IEEE; 2018. [DOI 10.1109/IW-
[20] Manuels MSD. « Analyses de laboratoire pour DRL.2018.8358214].
les maladies du sang ». [Online]. Available: [33] Sheng B, Zhou M, Hu M, Li Q, Sun L et Wen
https://www.msdmanuals.com/fr/accueil/troubles- Y. A blood cell dataset for lymphoma classifi-
du-sang/symptomes-et-diagnostic-des-troubles- cation using faster R-CNN. Biotechnol Biotech-
sanguins/analyses-de-laboratoire-pour-les- nol Equipment. 2020; 34(1): 413‑20. [DOI
maladies-du-sang. 10.1080/13102818.2020.1765871].
[34] Agrawal R, Satapathy S, Bagla G, Rajakumar K, dir. ence on Applied System Innovation (ICASI), 2024, p.
Detection of white blood cell cancer using image 107‑109. doi: 10.1109/ICASI60819.2024.10547902.
processing. International Conference on Vision
Towards Emerging Trends in Communication and
Networking (ViTECoN); 30-31 mars 2019; Vellore,
India. Vellore: IEEE; 2019. [DOI 10.1109/ViTE-
CoN.2019.8899602].
[35] Negm AS, Hassan OA, Kandil AH. A decision
support system for Acute Leukaemia classifica-
tion based on digital microscopic images. Alexan-
dria Engineering J. 2018; 57(4): 2319‑32. [DOI
10.1016/j.aej.2017.08.025].
[36] Sharma V, Rathore A, Vyas G. Detection of sickle
cell anemia and thalassemia causing abnormalities
in thin smear of human blood sample using im-
age processing. International Conference on In-
vention Computation Technologies (ICICT); 26-7;
Coimbatoire, India. Coimbatoire: Med; 2016. [DOI
10.1109/INVENTIVE.2016.7830136].
[37] Xu M, Papageorgiou DP, Abidi SZ, Dao M, Zhao H
et Karniadakis GE. A deep convolutional neural net-
work for classification of red blood cells in sickle cell
anemia. PLoS Comput Biol. 2017; 13(10): e1005746.
[DOI 10.1371/journal.pcbi.1005746].
[38] Mohamad A, Hamzah R, Mokhtar A, Sathar J,
dir. Sickle cell disease verification via sobel
edge algorithms for image processing. Interna-
tional Conference on Engineering Engineering
Technology and Technopreneurship (ICE2T); Kuala
Lumpur, Malaisie. 2017. Research Gate. [DOI
10.1109/ICE2T.2017.8215994].
[39] Zhang M, Li X, Xu M, Li Q. Automated semantic
segmentation of red blood cells for sickle cell dis-
ease. IEEE J Biomed. Health Inform. 2020; 24(11):
3095‑102. [DOI 10.1109/JBHI.2020.3000484].
[40] Alzubaidi L, Fadhel MA, Al-Shamma O, Zhang J,
Duan Y. Deep learning models for classification of
red blood cells in microscopy images to aid in sickle
cell anemia diagnosis. Electronics. 2020; 9(3): 427.
[DOI 10.3390/electronics9030427].
[41] Chy T, Rahaman M. A comparative analysis by KNN,
SVM, ELM classification to detect sickle cell ane-
mia. International Conference on Robotics, Elec-
trical and Signal Processing Techniques (ICREST);
Dhaka, Bangladesh. Dhaka: IEEE; 2019. [DOI
10.1109/ICREST.2019.8644410].
[42] Aliyu H, Razak M, Sudirman R, Ramli N. A deep
learning AlexNet model for classification of red
blood cells in sickle cell anemia. IAES IJ-AI. 2020;
9(2): 221-8. [DOI 10.11591/ijai.v9.i2.pp221-228].
[43] De Haan K, Ceylan Koydemir H, Rivenson Y, Tseng
D, Van Dyne E, Bakic L et al. Automated screening
of sickle cells using a smartphone-based microscope
and deep learning. Digit Med. 2020; 3(1): 76. [DOI
10.1038/s41746-020-0282-y].
[44] B. F. Bio Nigan, A. G. Zohoun, A. D. Kora. «
White Blood Cells Recognition and Classification
using Convolutional Neural Network », 2nd Interna-
tional Conference on Applied Artificial Intelligence
and Computing (ICAAIC), 2023, p. 145‑150. doi:
10.1109/ICAAIC56838.2023.10140293.
[45] B. F. Bio Nigan, A. G. Zohoun, A. D. Kora. « Auto-
matic Detection of Sickle Cell Disease, Elliptocyto-
sis and Schizocytosis », 10th International Confer-