Handwritten trip sheet to digital conversion using OCR

Neural Networks have promising potential at recognizing handwritten texts. In this paper, neural networks have been utilized for converting the handwritten trip sheet into digital format. A trip sheet is a written record containing the details of a trip. It is always easier to maintain records in a digitized format rather than in a physical form where there are chances of the documents being lost or damaged. Converting them into digital format solves all these problems making it easier & safer to organize and work on the data. The model proposed here is based on a combination of EasyOCR and Regular Expressions that help convert the data on a trip sheet into a CSV format. The proposed model works by splitting the image into various boxes by using the intersection of the lines and tries to recognize the characters present in that space separately. All of the recognized texts are then stored individually and can be later converted into various formats. The proposed model was tested on many handwritten trip sheets that provided favorable results with 85% accuracy.