An Improved Pathological Brain Detection System Based on Two-Dimensional PCA and Evolutionary Extreme Learning...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Machine Learning

Background:

• Pathological brain detection systems (PBDSs) have advanced, yet require improved accuracy for clinical use.
• Existing PBDSs often struggle to meet the demands of real-world diagnostic scenarios.

Purpose of the Study:

• To propose an efficient pathological brain detection system (PBDS) using MR images.
• To significantly enhance the accuracy of pathological brain detection compared to existing methods.

Main Methods:

• Image enhancement using contrast limited adaptive histogram equalization (CLAHE).
• Feature extraction via two-dimensional PCA (2DPCA) and PCA+LDA.
• Classification using a novel MDE-ELM algorithm combining modified differential evolution (MDE) and extreme learning machine (ELM) for optimizing neural network parameters.

Main Results:

• The proposed MDE-ELM system demonstrated superior performance on three standard datasets.
• Achieved higher accuracy in classifying pathological versus healthy MR images compared to conventional algorithms.
• The MDE-ELM classifier resulted in a more compact network architecture.

Conclusions:

• The proposed efficient PBDS based on MR images significantly improves pathological brain detection accuracy.
• The MDE-ELM algorithm offers a promising approach for accurate and efficient brain image analysis.
• This system holds potential for real-world diagnostic applications in neurology.