Handwritten Text Recognition Using Machine Learning
Character and have different shapes in this paper and it is really hard to recognize certain shapes. There are some confounding characters due to variations of shapes and there are very strong chances for misclassification or misrecognition. Neural networks are generally applied to recognition areas that are also handwritten. When using the Neural Networks the trained database is used and then tested on handwritten digits. In this work, the aim is to recognize the digits and special symbols the digits and special symbols by collecting the various text from the database of KAGGLE and MNIST. The database is collection of images of the handwritten text. The images are pre-processed, segmented and features are extracted. The subset of extracted features used for training the classifier. The performance of the classifier is tested for accuracy. Conventional l Neural Network and capsule Neural Network are used as classifier and their results are compared on digits and special symbol is recognition. Multiple learning techniques based on neural network for the handwritten recognition for characters , digits and special symbols ,and also new accuracy level for MNIST and KAGGLE dataset .The framework involves four primary levels are pre-processing , segmentation, feature extraction and CNN .This study improves the recognition accuracy more than 95% for the handwritten recognition for characters and digits. The database is collected and recognition the handwritten accuracy.