Offline Marathi Character To Text Conversion

India is a country with rich literature and historical importance. Most of this literature is written in Devanagari script. It is important to preserve this literature and pass the knowledge to the other generation. Marathi is the 15th most spoken language in the world. The most challenging research area is handwritten optical character recognition, because of its difficulty in segmenting the character that grows due to modifiers and compound characters due to Marathi Script Marathi handwritten character to text conversion is a system which provides the user the Optical Character Recognition functions on Marathi handwriting. Offline Marathi conversion character recognition system has mainly four stages: Pre-processing, segmentation, recognition and documentation. It uses the Convolutional Neural Network to train the dataset and uses Image Processing techniques to recognize user image via local storage. The good image quality is taken as input by the user to do further processing to get best results. This system uses python programming language and different hardware platforms, software platforms to make the system as a whole. The implemented system has an accuracy of 99.2%.