Data Extraction From Invoices Using Computer Vision

Management of invoices and maintaining their records for further processing sometimes it is hectic and buying specially developed software is not worth for small enterprises. The majority of the businesses has similar requirements and most of them use the traditional management system like recording data manually and maintaining hardcopies, a result, it consumes a lot of time as well as space. The proposed system is a web-based application specially built for small businesses like retailers and wholesalers. It is a standard business application made according to the standard requirements of businesses. The prime objective is to make data available for the user so that the user can access it anytime from anywhere and can modify it if need. Considering availability and accessibility this web-based application will help to achieve objectives. Our system focuses on the extraction of invoice data and segregates it into different parts like Vendor name, Invoice number, item name, and quantity. Our system tries to find out a better Management solution for auditing. We have initially focus on image invoices. The proposed solution follows three main steps preprocessing with the help of different OpenCV techniques and functions, then the next step is data extraction with the help of OCR technology and post-processing with the help of RegEx technology for better accuracy. The proposed system is webbased applications, specially build for small businesses like retailers and wholesalers. It is a standard business application made according to the standard requirements of businesses. The prime objective is to make data available for the user so that the user can access it anytime from anywhere and can modify it if need. Considering availability and accessibility, this web-based application will help to achieve objectives.