Automated Identity Document Classification

A model to classify documents into their categories and use them as per needs as in all the industrial documentation is needed to keep track of their employees documents and it is a frequent happening where all the documents get mixed up into each other and result in a lot of trouble and delay and leads to incurring a cost to the organization; the model presented in this paper offers a solution to this issue. A novel algorithm is implemented to tune the hyperparameter using a CNN Model as its basis using hence, created parameters to classify documents into their categories and making it easier to maintain we achieved an 81% average accuracy score which might be because of different countries belonging the datasets which have different structures as well as we have used one hot encoding which gives it a nice scope to be directly used to scale up and use in larger scale scenarios. Further, we implemented different machine learning models which provided us with a reliable accuracy of 94 percent which also proves that in a few cases, machine learning can outperform deep learning. Also it further gives us a great implacable model for real-life use cases.