OCR System:

1. Developed an Optical Character Recognition microservice capable of extracting useful details from images/receipts. The system was capable of rectifying the orientation of the receipts, split the images in case they contained multiple receipts.

2. The system was capable of extracting billing date, amount, currency, merchant name and categorize the receipts. I Implemented Support Vector Machine(SVM) technique to categorize the receipts and was able to achieve an accuracy of 82%. The overall accuracy of the systemwas 87%. The system was scaled to support multiple languages (English, French and German).

Worked on development of an Expense Management System which :

1. Has the capability to extract details out of receipts attached to it.

2. Split an image/pdf having multiple receipt to individual images/receipts.

Worked on :