Hi
This project can be split into several parts. Each one includes some Computer vision, machine learning or neural network components. Let me count them here.
1) object detection and isolation. For example detection of ID card and masking out background noise. Involves Support Vector Machines (SVM) and Convolutional Neural Network (CNN)
2) Object type recognition. For example recognition of ID card and type of ID card(e.g. Medicare card). Involves Recurrent CNN. Will need specific dataset for documents you target.
3) Form fields recognition. For example face picture or signature or DoB field. Involves CNN with specific dataset for passport recognition. This includes OCR.
4) Output Layout Form template. Here I include ODBC database, Form generator, Mobile and Desktop apps for many platforms (Linux, Windows, Mac, Android, IOS) for and other minor things
I will expect from you good explanation of details and training dataset for all neural networks. Also please indicate if it involves any other language rather than English
As for my experience. I worked on a projects that included recognition of receipts and ID cards. It was developed in the same pattern. For example goal of one project was to recognize ID and business cards, OCR, determine field types and store into database.
Thanks for your attention and waiting for your response
Archil