الفهرس 
Chapter 1 IntroductionOnly 14 pages are availabe for public view
 |  | from | 133 |  |  |
| | | from | 133 | | |
Abstract
Detection of cancer cells and healthy cells is one of the major concerns for saving human life, especially those with leukemia. Attempts were made in this thesis to detect and classify acute lymphoblastic leukemia from microscopic blood images. The used dataset included microscopic blood smear images from patients with suspected acute lymphoblastic leukemia (ALL), whose blood samples were staged and stained by highly skilled laboratory personnel. This dataset was divided into two categories: benign and malignant in the case of binary classification and divided into four categories: Benign, Early Pre-B, Pre-B and Pro-B in the case of multi classification.•First, the image is enhanced using adaptive threshold.•Second, feature extraction was performed using feature extraction methods.• Third, the number of features has been reduced by using the grey wolf for binary classification. Enhanced Grey Wolf Optimization (EGWO) based on k-means clustering algorithm was used to reduce the number of features for multi classification.• Finally, acute lymphoblastic leukemia was classified into:•Benign and malignant in the case of binary classification•Benign, Early Pre-B, Pre-B, and Pro-B in the case of multi classification by using K-Nearest Neighbors (KNN) classifiers, Support Vector Machine (SVM), Naive Bayes (NB), and Random Forest (RF) classifiers. After using the grey wolf optimization technique in the feature selection process, it has been shown to provide a good efficiency in all cases. This thesis can improve the performance and reduce the number of features after using the technique of grey wolf optimization in binary classification and enhanced grey wolf optimization in multi classification, hence this is the first time that (GWO) and (EGWO) are used in the detection and classification of acute lymphoblastic leukemia (ALL) into subtypes as feature selection techniques. The results shown that the proposed method can be used as a tool for diagnosing acute lymphoblastic leukemia and its subtypes, which will certainly help pathologists.