Receipt recognition project

Goal

To build an application that automatically extracts information from pictures of receipts.

Current progress

Working on detect regions based on a sample.

Dependencies

• python3
• cv2 (OpenCV 3.4.4)
• numpy
• Tesseract OCR

Usage

Download

$ git clone https://github.com/nv-quan/receipt-recognition.git
$ cd receipt-recognition

Config

In ./src/config.json,

• input file: directory of input picture
• input mode: single to process one file at a time, folder to process all file in one folder
• input folder: directory of input folder if input mode is folder
• run mode: normal for normal mode, debug for debug

In ./docsample/sample.json, sample is the directory of the sample image

Register correct regions

$ python3 ./src/register.py

Drag in the image to detect sample region, type the ids for them.

Run

$ python3 ./src/main.py