The W Lovers Bill.com Challege

In this project we use python OCR libraries to read text from receipts so that we can match them to a database of manually entered receipt data. We think this to be robust model for Bill.com to use.

Comment

Inspiration

We were inspired by Bill.Com's need for receipt matching in a non tedious form. We felt our model would be the perfect fit to achieve their goals.

What it does

It uses OCR to create bounding boxes around the text in the image and extract the key features. We use the key features to match the extracted OCR data to the manually entered receipt information data table. Eliminating the need for a human to scroll and search for the correct transaction.

How we built it

We built it using python as well as a python library called pytesseract. We also embedded a free online OCR API to extract the OCR data quickly.

Challenges we ran into

The challenges we ran into dealt with successfully extracting structured data from the OCR output and building a function that can take that data and match it with the proper row transaction from the data table.

Accomplishments that we're proud of

We were proud that we were able to successfully extract text from the receipt images as well as create bounding boxes around the key features.

What we learned

We learned how to use and implement OCR models as well as how to use different APIs.

What's next for Bill.com Challege

We will continue to grow our knowledge on the subject and build more models relating to OCR.

Built With

Share this project:

Updates