A Hemolysis Image Detection Method Based on GAN-CNN-ELM.

Area of Science:

• Medical Imaging Analysis
• Machine Learning in Diagnostics
• Computational Pathology

Background:

• Manual hemolysis testing is costly and labor-intensive.
• Objective and automated methods are needed for accurate hemolysis detection.
• Image analysis offers a promising avenue for hemolysis assessment.

Purpose of the Study:

• To develop an automated hemolysis image detection method.
• To improve the accuracy and efficiency of hemolysis testing.
• To leverage deep learning for enhanced diagnostic capabilities.

Main Methods:

• Image and data enhancement techniques were applied.
• Generative Adversarial Networks (GANs) were used for data augmentation.
• Convolutional Neural Networks (CNNs) extracted image features.
• Extreme Learning Machine (ELM) optimized the classification model.

Main Results:

• The proposed GAN-CNN-ELM model achieved a high accuracy rate of 98.91%.
• The model demonstrated superior performance in both accuracy and speed compared to other tested models (GAN-CNN, GAN-ELM, GAN-SVM, CNN-ELM).
• The method effectively identifies and categorizes hemolysis images.

Conclusions:

• The GAN-CNN-ELM model provides an effective and efficient solution for automated hemolysis image detection.
• This approach offers a significant improvement over traditional manual testing methods.
• The study highlights the potential of integrated deep learning models in medical diagnostics.