Latent Space Representational Learning of Deep Features for Acute Lymphoblastic Leukemia Diagnosis

Acute Lymphoblastic Leukemia (ALL) is a fatal malignancy that featured by the abnormal increase of immature lymphocytes in blood or bone marrow. Early prognosis ALL indispensable for effectual remediation this disease. Initial screening conducted through manual examination stained smear microscopic images, process which time-consuming and prone to errors. Therefore, many deep learning-based computer-aided diagnosis (CAD) systems have been established automatically diagnose ALL. This paper proposes novel hybrid learning system images. The introduced integrates proficiency autoencoder networks feature representational latent space with superior extraction capability standard pretrained convolutional neural (CNNs) identify existence smears. An augmented set image features are formed from extracted GoogleNet Inception-v3 CNNs dataset A sparse network designed create an abstract significant enlarged set. used perform classification using Support Vector Machine (SVM) classifier. obtained results show improve performance proposed over original features. Moreover, various sizes evaluated. retrieved reveal superiorly compete state art.