Hybrid deep layered network model based on multi-scale feature extraction and deep feature optimization for acute...

Area of Science:

• Hematology
• Medical Diagnostics
• Artificial Intelligence in Medicine

Background:

• Acute lymphoblastic leukemia (ALL) is a common childhood hematological malignancy.
• Early ALL diagnosis is crucial for effective cancer treatment and patient management.
• Traditional ALL detection methods are time-consuming and rely on subjective expert interpretation.

Purpose of the Study:

• To comparatively analyze the performance of a novel hybrid deep learning model for ALL diagnosis.
• To develop an automated system for accurate detection and classification of ALL cells.
• To overcome the limitations of traditional diagnostic methods in terms of speed and objectivity.

Main Methods:

• A hybrid deep learning model integrating Xception architecture for feature extraction and Extreme Gradient Boosting (XGBoost) for classification was developed.
• Blood cell images were preprocessed using a center-based cropping strategy to remove irrelevant areas.
• The dataset was divided into training, validation, and test sets for robust model evaluation.

Main Results:

• The proposed hybrid model achieved a high accuracy of 98.88% in detecting ALL.
• The model demonstrated superior performance compared to existing hybrid models in ALL detection.
• Automated feature extraction and optimization led to enhanced classification accuracy.

Conclusions:

• The developed hybrid deep learning model offers a highly accurate and efficient solution for ALL diagnosis.
• This AI-driven approach has the potential to significantly improve the speed and reliability of medical decision-making in ALL cases.
• The study highlights the effectiveness of combining deep learning architectures with advanced classifiers for hematological disease detection.