Cloud-Enhanced Machine Learning for Handwritten Character Recognition in Dementia Patients

Abstract

The study addresses the challenge dementia patients face in recognizing handwritten characters by developing a cloud-integrated system that uses a multilayer neural network for character recognition. The system involves four main steps: preprocessing (noise reduction and normalization), segmentation (extracting characters from scanned pages), feature extraction (using a modified zone-based method), and recognition. The extracted features, represented as pixel value vectors, are classified using four machine learning algorithms—support vector machine with RBF, random forest, linear SVM, and logistic regression. The random forest algorithm performs best with an accuracy of 89%. Cloud technology enhances the system's scalability, allowing for real-time processing and remote access, beneficial for dementia care.