Bank form and signature checker
Recognise scanned bank forms and check that every required signature is there.
The brief
Branch staff check thousands of scanned forms by eye, and a missing signature delays the customer for days. Train an image model that recognises the type of form and checks whether each signature box is signed, sending doubtful cases to a person. Good looks like: high accuracy on form type, few missed blank signatures, and a review queue for unsure cases.
What you will build
- Build a labelled set of form images (public and synthetic)
- Classify the form type with a pre-trained image model
- Detect signature regions and decide signed or blank
- Measure missed blanks and false alarms
- Show results with a confidence-based review queue
Free, open data to build it with
- FUNSD: form understanding in noisy scanned documents ↗ FUNSD
- Handwritten signatures ↗ Kaggle
Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.

