Abstract: Computers and phones may be more common than ever, but most people still prefer the traditional way of writing with ink on paper. People in the rural parts of India are mostly comfortable with the pen and paper way of going about their work. But with rapid technology advancements, everything has gone digital from Aadhar card forms to Birth Certificates. Despite this easy availability of a vast number of technical writing tools, many people choose to take their notes traditionally in the written manner in the language they are comfortable with, which is usually Hindi. Our work is on word recognition of handwritten Hindi characters and its implementation on handwritten forms. Our paper introduces an end-to-end word spotting system for the Hindi language using Segmentation based approaches. Our proposed architecture implements an end-to-end strategy that recognizes handwritten Hindi words from printed forms and is translated into English. Hence, handwriting recognition and translation interpret the Hindi handwritten input from various handwritten sources, such as paper documents, forms, into digital form translated into English. A form recognition system handles the formatting, performs correct segmentation into characters, and detects the Hindi words, which are then translated into English and shown on the form. The computational study of people’s opinions, sentiments expressed is termed as sentiment analysis which is also known as opinion mining. For the feedback forms, sentiment analysis is performed using Random Forest algorithms and NLTK libraries like Porter stemmer and Stop words are used giving an accuracy of 88%.

Keywords : OCR , NLP , Image Processing , Sentiment analysis

PDF | DOI: 10.17148/IJARCCE.2022.11236

Open chat
Chat with IJARCCE