Directional mutation and crossover boosted ant colony optimization with application to COVID-19 X-ray image...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computer Vision

Background:

• Accurate segmentation of Coronavirus Disease 2019 (COVID-19) X-ray images is crucial for diagnosis.
• Existing segmentation methods may face challenges with accuracy and convergence speed.

Purpose of the Study:

• To develop an enhanced multilevel image segmentation method for COVID-19 X-ray analysis.
• To introduce an improved swarm intelligence algorithm (XMACO) for superior image segmentation.

Main Methods:

• An improved ant colony optimization algorithm (XMACO) with directional crossover (DX) and directional mutation (DM) strategies was developed.
• A multilevel image segmentation model (MIS-XMACO) was designed using 2D histograms, 2D Kapur's entropy, and a nonlocal mean strategy.
• The XMACO algorithm was benchmarked against IEEE CEC2014 and IEEE CEC2017 functions, and the MIS-XMACO model was applied to COVID-19 X-ray segmentation.

Main Results:

• XMACO demonstrated faster convergence speed, higher accuracy, and effective avoidance of local optima compared to other algorithms.
• The MIS-XMACO model achieved more stable and superior segmentation results on COVID-19 X-ray images across various threshold levels.
• Experimental validation confirmed the effectiveness and superiority of the proposed model over existing segmentation techniques.

Conclusions:

• The proposed XMACO algorithm and MIS-XMACO model offer significant improvements in COVID-19 X-ray image segmentation.
• This advancement has the potential to enhance the accuracy and efficiency of COVID-19 diagnosis through medical image analysis.