AI-driven genetic algorithm-optimized lung segmentation for precision in early lung cancer diagnosis

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• Lung cancer is a leading cause of mortality globally.
• Accurate lung segmentation is crucial for early diagnosis and treatment planning.
• Existing diagnostic tools require improvement in efficiency and accuracy.

Purpose of the Study:

• To develop an advanced AI-driven framework for precise lung segmentation in early lung cancer diagnosis.
• To optimize the framework using genetic algorithms for improved accuracy and reduced computational complexity.
• To enable efficient and accurate lung cancer diagnosis, particularly in resource-constrained environments.

Main Methods:

• Utilized the UNET3+ architecture integrated with multi-scale feature extraction.
• Employed genetic algorithms to optimize neural network configurations and identify optimal parameters.
• Conducted extensive experiments on publicly available lung segmentation datasets.

Main Results:

• Achieved a Dice Similarity Coefficient of 99.17%, demonstrating superior segmentation accuracy.
• Reduced model parameters by 74% compared to the baseline UNET3+ model, significantly lowering computational cost.
• Validated the framework's effectiveness on diverse lung segmentation datasets.

Conclusions:

• The AI-driven framework offers a highly accurate and computationally efficient solution for lung segmentation.
• The optimized model is suitable for deployment in resource-limited settings and point-of-care diagnostic devices.
• This approach highlights the potential of AI to enhance early lung cancer detection and reduce healthcare disparities.