Medical Image Processing of Alzheimer

A progressive disease Alzheimer is the leading source of dementia in the world, and its frequency continues to increase, in part due to the aging of the world’s population. The process of this disease is naturally characterized by two characteristic pathologies. The diagnosis is constructed on clinical presentation and fluid and imaging biomarkers that meet several criteria. There was a requirement to automatically diagnose definite diseases based on medical images and subjects. This helps doctors and radiologists take further steps to treat the disease. Alzheimer’s disease was chosen for this purpose. Alzheimer’s disease is the leading cause of dementia and forgetfulness. Major cause of Alzheimer’s disease is by the atrophy of a particular brain region and the death of brain cells. MRI scans give this information, but the areas of atrophy vary from person to person, making diagnosis a bit more difficult and are often misdiagnosed by doctors and radiologists. Provided by KAGGLE with more than 6400 subjects. Kernel convolution and neural networks are combined to create convolutional neural networks (CNN) . Kernel convolution is a technique for recognizing and segmenting images based on features using filters. A neural network represents a single classifier and consists of neurons loosely based on human brain neurons that are interconnected by weights, have different biases, and are activated by some activation function. Using a convolutional neural network solves the problem with a minimal error rate. The primary purpose of this theory is to demonstrate a novel method for classifying Alzheimer’s disease (AD) built with TensorFlow and convolutional neural networks (TF and CNN). The network has three layers: a convolutional layer for removing AD characteristics, a flattening layer for bringing proportions down, and two fully connected layers for categorizing the withdrawn features. TensorFlow’s primary goal is to build computational graphs. Two of his main contributions were made to improve the performance of classification: data enrichment and multi-optimizer. Data augmentation helps reduce overfitting and improve model performance. Images in the training dataset are enhanced by normalization, rotation, and cropping.