Research and Application of Deep Learning Models for Detecting Brain Abnormalities Based on CT Images

This paper studies a method to improve the accuracy of a deep learning model for detecting brain abnormalities based on computed tomography images. The process begins with image preprocessing using the Histogram Equation algorithm and Gabor filter. Then, features are extracted from the fully connected layer of the AlexNet model. To optimize feature extraction, we use the MIFS algorithm to identify the most important features. Finally, the SVM machine learning model is deployed to detect brain abnormalities. To confirm the effectiveness of the proposed model, we compare it with other popular deep learning models, including AlexNet. Experimental results show that our method achieves an accuracy of over 93%, higher than pure deep learning models.