The advancements in machine learning and artificial intelligence have touched all the traditional professions. Accountancy is changing and developing as a result of technology and it’s giving rise to a new domain called Accounting Engineering. Invoice processing is a part of accounting jobs that involves a human finding and processingthe information in the invoices. After reading the required data from the invoices, the accountants classify the invoices into various accounts. In this project, I extracted the data from invoice images and explored a few classifiers that will consume this data and categorize the invoices into the target accounts.I have experimented with support vector machine, logistic regression, recurrent neural networks and random forest models, along with text encodings like TF-IDF and count vector. With limited availability of data, the maximum accuracy attained by the classifiers was 81%, around 22% improvement over the baseline. With access to more trainingdata, these methods could prove to be a promising platform for further research.Keywords: Natural language processing, automation of accounting processes, text encoding, word embeddings, Recurrent Neural Networks, LSTM, Random Forest, SVM, regression