Enhancing Alzheimer Disease Diagnosis: Integrating Gabor Convolutional Neural Network with Conventional CNNs

Alzheimer’s disease (AD) presents a formidable challenge in healthcare, characterized by its progressive nature and the absence of definitive diagnostic markers. This study introduces an innovative method to enhance AD detection by integrating advanced technology with established techniques. We have developed a hybrid model that combines Gabor Convolutional Neural Networks (GCNN) with traditional Convolutional Neural Networks (CNNs) to improve feature extraction from MRI images, thereby achieving a more precise diagnosis of AD. The efficacy of our model is evidenced by its high-performance metrics: an accuracy of $98.58 \%$, a sensitivity of $\mathbf{9 8 . 1 7 \%}$, a precision of $97.11 \%$, and robust empirical support. This approach benefits from the use of augmented data and the unique capabilities of GCNN. Comprehensive evaluations, including confusion matrix analysis and ROC curve assessments, confirm the model’s proficiency in accurately classifying various stages of dementia. By enhancing diagnostic accuracy, this innovative approach holds the potential to improve patient outcomes, facilitate better care, and enable more timely interventions in the treatment of Alzheimer’s disease.