Pipelined deep learning architecture for the detection of Alzheimer's disease

The progressive neurodegenerative disease in the human brain causes Alzheimer's disease (AD).The earlier detection helps to slowdown the progression of AD using continuous medical support system. The research work aims in the detection of Alzheimer's disease (AD) using brain images. The AD images are detected using fusion- based deep learning method. This works also adopts Pipelined LeNet (PLN) architecture. The brain Magnetic Resonance Imaging (MRI) images obtained are resized in the preprocessing stage and the internal regions are enhanced using the image fusion method. Therefore, the proposed AD detection system produces high classification rate even using low resolution brain MRI images. The ternary features are computed from the fused image and these features are classified by the proposed PLN architecture for identification of Alzheimer's brain image. High efficient and high speed novel PLN architecture is proposed in this AD detection system using brain MRI images due to the consumption time of classification and requirement of brain image dataset by conventional methods. Evaluation of the work is carried out using Kaggle Alzheimer's Classification Dataset (KACD). Performance of the AD image detection system in terms of sensitivity (Se), specificity (Sp), precision (Pr), and accuracy (Acc) are evaluated. The proposed AD image detection system obtains 99.9% Se, 99.8% Sp, 99.8% Pr and 99.5% Acc using PLN architecture. The proposed AD detection system using PLN architecture consumed 0.65 ms as an average execution span. By comparing the execution span of all other similar methods, the proposed system consumes less execution span due to its Pipelined architecture. The experimental results of the image detection system are compared with similar works in the literatures.