Accueil > Résultats de la recherche > Suite de la publication

Convolutional Neural Network and decision support in medical imaging: case study of the recognition of blood cell subtypes

Abstract: Identifying and characterizing the patient's blood samples is indispen- sable in diagnostics of malignance suspicious. A painstaking and sometime sub-jective task are used in laboratories to manually classify white blood cells. Neural mathematical methods as deep learnings can be very useful in the automated recognition of four (4) subtypes of blood cells for medical application. The purpose of this study is to use deep learning for image recognition of the four (4) blood cell types and to enable it to tag them. These approaches therefore depend on convolutional neural networks. To do this, we have a dataset of blood cells with labels of the corresponding cell types. The elements of the database are the input of our convolution which is a simple mathematical tool that is widely used for image processing. These databases have allowed us to create learning models for image recognition, particularly of the blood cell type. Based on the fact that a deep neural network model is able to rec-ognize each element of a scene provided it has been trained for this purpose, this activity focused on carefully selecting the optimization parameters of the model. We evaluated the recognition performance and outputs learned by the networks in order to implement a neural image recognition model capable of distinguishing polynuclear cells (neutrophil and eosinophil) from those of mononuclear cells (lymphocyte and monocyte). The classification accuracy on the learning dataset is 97.39% and the validation accuracy is 97.77%. Images detection failure is very low.


Auteur(s) : Daouda DIOUF1, Djibril SECK2, Mountaga DIOP2 and Abdoulaye BA3 1Laboratoire de Traitement de l’Information, Ecole Supérieure Polytechnique, Université
Pages : 04-06
Année de publication : 2019
Revue : CEUR-WS.org/Vol-2647/paper10.pdf
N° de volume : vol-2647
Type : Article
Statut Editorial : Copyright © 2020 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). IREHI-2019: International Conference on rural and elderly health Informatics, Dakar, Sénégal, December 04-06, 2019
Mise en ligne par : DIOP Mountaga