the bone marrow producing many white blood cells (WBC). ALL is one of the most common types of cancer in children and adults. Doctors determine the treatment of leukemia according to its stages and its spread in the body. Doctors rely on analyzing blood samples under a microscope.
Pathologists face challenges, such as the similarity between infected and normal WBC in the early stages. Manual diagnosis is prone to errors, differences of opinion, and the lack of experienced pathologists compared to the number of patients. Thus, computer-assisted systems play an essential role in assisting pathologists in the early detection of ALL.
In this study, systems with high efficiency and high accuracy were developed to analyze the images of C-NMC 2019 and ALL-IDB2 datasets. In all proposed systems, blood micrographs were improved and then fed to the active contour method to extract WBC-only regions for further analysis by three CNN models (DenseNet121, ResNet50, and MobileNet). The first strategy for analyzing ALL images of the two.
extract the boundaries of WBC cells because it provides a smooth contour. A contour is a border to select WBC cells; it is a set of points subject to the interpolation process, which is polylinear to define the object’s boundaries.
Active contour is a process for obtaining deformable structures with constraints and forces in segmentation blood microscopy im- ages, which describe WBC cell boundaries and other features to form contour and curve parameters [ 35]. The curve models are determined by applying intrinsic and extrinsic forces through Equation (3) .
These forces are related to the curves of the images of microscopic cells. External energy is the combination of energy that controls the circumference of WBC cells within the image. Internal energy controls deformable changes [36]. GcvC=Z outside(I(x) m1)2dx+Z inside(I(x) m2)2dx+blength (C) (3) where Irefers to gray intensity, outside andinside refer to the regions outside and inside the contour.
inputted the medical field, which has high capabilities to identify many diseases, including distinguishing between normal and blasted blood cells. CNN automatically learns the hierarchy of spatial features from the input images. CNN has millions of parameters, biases, weights, and connections that can be learned and adapted while training the data.
The weights and parameters are adjusted by minimizing the difference between the actual and expected data through the backpropagation of the network. The main motivation of this work is to develop effective microscopic blood slide analysis models for the diagnosis of ALL.
The medical dataset lacks huge images, making it difficult to train the CNN network from scratch so that pre-trained CNN models can extract features from the C-NMC 2019 and ALL_IDB2 datasets. To improve the performance of the proposed systems, PCA was used to select only important features, combine features of CNN, and feed them to RF and XGBoost algorithms. The main major contributions of this.
models to produce a strong model. This process is called Boosting. When training the network, the weights of the variables predicted by the method (decision tree) are incorrectly increased [ 51]. The variables are sequentially fed to the next decision tree. Boosting is an ensemble learning method for creating strong classifiers from serially weak classifiers.
The first model is built to predict the training dataset, then, building the next predictive model to solve the errors in the previous model, and the process continues until the predicted training dataset is accurate [52]. This hybrid strategy has two techniques.
The first technique is shown in Figure 5, which goes through the following sequence: first, optimizing the images of the C-NMC 2019 and ALL-IDB2 datasets. Second, extracting WBC cells and isolating them from other major blood components by the active contour method and keeping them in a folder in a new dataset called ALL-IDB2-ROI. Third, the C-NMC 2019 and ALL-IDB2-ROI datasets are fed into.
arrow [ 8]. The few normal blood cells are discharged into the bloodstream which cannot supply the body’s organs with sufficient oxygen, so that the immune system will weaken and the blood clotting will be weak. ALL represents 25% of childhood cancers; 74% of leukemia cases in people under 20 have ALL.
The five-year survival rate for children under 14 is 91%, while people between the ages of 15 and 20 have a five-year survival rate of 75%. In all cases diagnosed early and treated, the ALL disease cannot return, meaning that children diagnosed with ALL, after five years, became healthy and recovered completely.
However, the recovery rate for adults five years after their injury is not high, ranging between 20–35% [ 9]. ALL spreads quickly if left undiagnosed and can lead to death within months. ALL is usually diagnosed by a complete blood count test, in which a doctor checks for certain clinical signs of leukemia [ 10]. Sometimes the doctor is unsatisfied with the symptoms of a complete blood count smear.
iro, A.B. Resistance to Tyrosine Kinase Inhibitors in Chronic Myeloid Leukemia—From Molecular Mechanisms to Clinical Relevance. Cancers 2021 , 13, 4820. [CrossRef] [PubMed] 51. Mandal, M.; Singh, P .K.; Ijaz, M.F.; Shafi, J.; Sarkar, R. A Tri-Stage Wrapper-Filter Feature Selection Framework for Disease Classification. Sensors 2021 ,21, 5571. [CrossRef] 52. Senan, E.M.; Abunadi, I.; Jadhav, M.E.;.
Score and Correlation Coefficient-Based Feature Selection for Predicting Heart Failure Diagnosis by Using Machine Learning Algorithms. Comput. Math. Methods Med. 2021 ,2021 , 8500314. [CrossRef]Diagnostics 2023 ,13, 1026 33 of 33 53. Mohammed, B.A.; Senan, E.M.; Alshammari, T.S.; Alreshidi, A.; Alayba, A.M.; Alazmi, M.; Alsagri, A.N.
Hybrid Techniques of Analyzing MRI Images for Early Diagnosis of Brain Tumours Based on Hybrid Features. Processes 2023 ,11, 212. [CrossRef] 54. Al-Hejri, A.M.; Al-Tam, R.M.; Fazea, M.; Sable, A.H.; Lee, S.; Al-antari, M.A. ETECADx: Ensemble Self-Attention Transformer Encoder for Breast Cancer Diagnosis Using Full-Field Digital X-ray Breast Images. Diagnostics 2023 ,13, 89. [CrossRef].