Deep and Hybrid Learning Techniques for Diagnosing Microscopic Blood Samples for Early Detection of White Blood Cell Diseases

The immune system is one of the most critical systems in humans that resists all diseases and protects body from viruses, bacteria, etc. White blood cells (WBCs) play an essential role system. To diagnose diseases, doctors analyze samples to characterize features WBCs. characteristics WBCs are determined based on chromatic, geometric, textural WBC nucleus. Manual diagnosis subject many errors differing opinions experts takes a long time; however, artificial intelligence techniques can help solve these challenges. Determining type using automatic helps hematologists identify different types diseases. This work aims overcome manual by developing automated for classifying microscopic sample datasets early detection Several proposed were used: first, neural network algorithms, such as networks (ANNs) feed-forward (FFNNs), applied dataset extracted hybrid method between two local binary pattern (LBP) gray-level co-occurrence matrix (GLCM). All algorithms attained superior accuracy diagnosis. Second, pre-trained convolutional (CNN) models AlexNet, ResNet-50, GoogLeNet, ResNet-18 exceptional results Third, technique was applied, consisting pair blocks: CNN block extracting deep SVM algorithm classification with efficiency. These named AlexNet SVM, ResNet-50 GoogLeNet SVM. achieved promising when diagnosing model 99.3%, precision 99.5%, sensitivity 99.25%, specificity 99.75%, AUC 99.99%.