Adaptive genetic algorithm based deep feature selector for cancer detection in lung histopathological images

Area of Science:

• Medical Imaging
• Computational Pathology
• Artificial Intelligence in Oncology

Background:

• Cancer poses a significant global health challenge, necessitating early detection and accurate classification for effective treatment.
• Histopathology images are vital for diagnosing cancer, determining disease stage, and guiding treatment decisions.
• Deep learning models, particularly convolutional neural networks, show promise in analyzing medical images for cancer detection.

Purpose of the Study:

• To develop and evaluate a novel deep learning model for accurate lung cancer classification using histopathology images.
• To enhance feature extraction and selection for improved diagnostic performance in lung cancer detection.
• To provide a robust and accurate automated system for lung cancer analysis.

Main Methods:

• A channel attention-enabled deep learning model was employed as a feature extractor.
• An adaptive Genetic Algorithm (GA) was utilized for feature selection, calculating fitness scores using a filter method.
• The GA-optimized feature vector was then classified using a K-nearest neighbors (KNN) classifier.

Main Results:

• The proposed method achieved a high overall accuracy of 99.75% on the LC25000 dataset.
• The integrated approach demonstrated effective feature extraction and selection for lung histopathological images.
• The study highlights the potential of combining deep learning with genetic algorithms for cancer classification.

Conclusions:

• The developed channel attention-enabled deep learning model with GA-based feature selection offers a highly accurate method for lung cancer classification.
• This approach shows significant promise for improving the accuracy and efficiency of automated lung cancer diagnosis from histopathology images.
• The findings contribute to the advancement of AI-driven tools in digital pathology and oncology.