Digitization of Physical Notes: A Comprehensive Approach Using OCR, CNN, RNN, and NMF

2.3 trillion gigabytes of data is generated every day but that's today's world for you. In this technological era where we can use technology in almost all sectors and regular and constant day-to-day activities, these huge digital copies and data are communicated and accessed. Hence even though handwritten texts have their purpose and application they are of no use in the digital era. Questioning the same problem, innovations and key-to-note digitization and topic identification of this huge digital database carry inestimable merits and value. Before fitting execution, we went on exploring different methods, algorithms, and systems with our own which lead us to many trials and errors, and mistakes. Ultimately we came up with a system integrating Topic Detection and Identification. Intending to integrate both processes into a single system, we examine various algorithms involving neural networks like RNN, CNN, and ANN, and methods such as Tesseract, KNN, and LSTM that are used for implementing OCR while techniques such as K means clustering, TF-IDF, LDA, and LINGO have been employed to perform topic detection and identification. Based on our study and results from various papers, we have decided to use CNN for OCR.