We use cookies to ensure that we give you the best experience on our website Learn more

Sep 12, 2018

Society of Hematologic Oncology Sixth Annual Meeting

06 / Computer aided Leukemia Detection using Microscopic Blood Image based Machine Learning "Convolutional Neural Network"







Background:standard morphologic diagnosis of leukemia by hematologists is done by examining patient peripheral blood and bone marrow under microscope, But, manual recognition is prone to errors due to variations as experience and tiredness, error is said to be between 30% - 40%, thus affects credibility ,hence, necessity for a robust automated system for leukemia detection that is not influenced by human variations.Machine learning is attracting tremendous interest in the biomedical field; improving sensitivity,specificity and decision-making objectivity of leukemia diagnosis.Objectives: testing computer aided systems that use microscopic images to recognize leukemia based on a pretrained deep convolutional neural network (CNN) which is an approach of machine learning algorithms.Material and method: dataset used was collected from Clinical Pathology lab,Assiut University Hospital,Egypt,containing binary balanced dataset:leukemic and normal blood, each sample composed of PB and BM smears images.We analyzed 1000 digital images ;500 images of leukemia (all types included) & 500 images of normal blood.1-Training & testing phase: testing different types of pre-trained (CNN) models ex.Alexnet, VGG16, VGG19, GoogLeNet, ResNet50, ResNet101, and Inception-v3.We divided images into two groups; 80% of images assigned as" training-set", each of the models trained on the images to extract features to differentiate leukemic and normal images, remaining 20% images were assigned as "testing-set" or so called "never seen" by the models .later we tested the types of leukemia against each other to measure sensitivity, specificity and accuracy.2-Evaluation of sensitivity/specificity/accuracy: evaluating ability of the models to detect target; based on sensitivity, specificity and accuracy of detection.Results: comparing the ability of models to distinguish leukemia vs normal images showed that VGG16 model had the lowest accuracy (92.13 %) and Inception-v3 model had the highest accuracy (99.98%).concerning sensitivity and specificity, the Inception-v3 can detect all leukemic cases of ALL, AML, CML, and CLL with 100% sensitivity; and specificity of 97.561%, 93.023%, 99.98%, and 95.238% respectively.Conclusion: inception-v3 model outperforms other pre-trained (CNN) in diagnosis of leukemia ,this model of Machine learning algorithm can be used in the context of lab diagnosis acting as a second opinion after manual evaluation of leukemia.

Discover over 20,000 new abstracts, posters and presentations from leading academic conferences every month. Stay on top of the latest findings, methodologies and discussions happening in your research field around the world.



Follow us

© Copyright 2019 Morressier GmbH. All rights reserved.

© Copyright 2019 Morressier GmbH.
All rights reserved.