The open source extract transaction infomation by using OCR.

Overview

Transaction OCR

Mã nguồn trích xuất thông tin transaction từ file scaned pdf, ở đây tôi lựa chọn tài liệu sao kê công khai của Thuy Tien. Mã nguồn có thể ứng dụng để giải quyết bài toán liên quan đến trích xuất thông tin văn bản từ hình ảnh (OCR - Optical Character Recognition) có cấu trúc nội dung xác định và với độ dài các dòng thông tin (row) bất kì như thông tin giao dịch, hóa đơn mua hàng,... Mã nguồn lựa chọn Cloud Vision API đại diện cho OCR model để có được độ chính xác cao, hoặc bạn có thể sử dụng model có sẵn như Vietocr hoặc có thể tự build custom OCR tiếng Việt từ clovaai: text-detectiontext-recognition) mà tôi cho là khá tốt.

Getting Started

Dependency

git clone https://github.com/hungtooc/transaction_ocr.git

pip install -r requirements.txt

1. Repair data input

1.1 Download raw data

1.2 Convert pdf files to image

PDF password: [email protected]

python tools/pdf-to-images.py --pdf-password [email protected]
usage: pdf-to-images.py [-h] [--pdf-dir PDF_DIR] [--output-dir OUTPUT_DIR] [--pdf-password PDF_PASSWORD] [--from-page-no FROM_PAGE_NO] [--to-page-no TO_PAGE_NO] [--fix-page-number FIX_PAGE_NUMBER]

optional arguments:
  -h, --help            show this help message and exit
  --pdf-dir PDF_DIR     dir to pdf files
  --output-dir OUTPUT_DIR
                        dir to save images
  --pdf-password PDF_PASSWORD
                        pdf password
  --from-page-no FROM_PAGE_NO
                        extra image from page
  --to-page-no TO_PAGE_NO
                        extra image to page
  --fix-page-number FIX_PAGE_NUMBER
                        fix page number (page_no += fix_page_number)

2. Extract transaction information

The source perform the basic steps to extract transaction information, you may want to add additional processing to optimize the source code in lines marked #todo.

python run.py 
usage: run.py [-h] [--image-dir IMAGE_DIR] [--output-respone-dir OUTPUT_RESPONE_DIR] [--output-content-dir OUTPUT_CONTENT_DIR] [--processed-log-file PROCESSED_LOG_FILE]

optional arguments:
  -h, --help            show this help message and exit
  --image-dir IMAGE_DIR
                        dir to images
  --output-respone-dir OUTPUT_RESPONE_DIR
                        dir to save api respone
  --output-content-dir OUTPUT_CONTENT_DIR
                        dir to save transaction content
  --processed-log-file PROCESSED_LOG_FILE
                        path to log file

File run.py perform 7 main stages:

  • Step 1. Find header & footer.
  • Step 2. Re-rotate image based on header-corner.
  • Step 3. Clean image.
  • Step 4. Call request google-ocr api. (include:text-detection & text-recognition
  • Step 5. Detect transaction line.
    processing-step-boder

  • Step 6. Classify transaction content each line & each content type.
    read-transactions-border

  • Step 7. Save transactions content to csv.
TNX Date Doc No Debit Credit Balance Transaction in detail (note)
13/10/2020 5091.55821 100.000 586062.131020.075756.Ung ho mien trung FT20287151644070 page_1
13/10/2020 5091.56080 1.000.000 586279.131020.075829.Ung ho dong bao mien Trung FT20287592192480 page_1
13/10/2020 5091.56138 200.000 219987.131020.075839.Trinh Thi Thu Thuy chuyen tien ung ho mien Trung page_1
13/10/2020 5091.56155 100.000 586295.131020.075826.UH mien trung FT20287432289640 page_1
13/10/2020 5078.68388 500.000 MBVCB.807033343.PHAM THUY TRANG chuyen tien ung ho tu thien.CT tu 0561000606153 PHAM THUY TRANG toi 0181003469746 TRAN THI THUY TIEN page_1
13/10/2020 5091.56261 1.000.000 184997.131020.075853.Em gui giup do ba con vung lu page_1
13/10/2020 5078.68496 200.000 MBVCB.807033583.Ung ho mien trung.CT tu 0051000531310 HUYNH THI NHU Y toi 0181003469746 TRAN THI THUY TIEN page_1
13/10/2020 5078.68526 100.000 MBVCB.807033514.ung ho mien trung.CT tu 0481000903279 NGUYEN THI HUONG AN toi 0181003469746 TRAN THI THUY TIEN page_1
13/10/2020 5091.56381 100.000 479592.131020.075909.ho tro mien trung page_1
13/10/2020 5078.68537 500.000 MBVCB.807034561.Ung ho Mien trung.CT tu 0721000588146 LE THI HONG DIEM toi 0181003469746 TRAN THI THUY TIEN page_1
13/10/2020 5091.56405 200.000 292363.131020.075845.Ngan hang TMCP Ngoai Thuong Viet Nam 0181003469746 LUC NGHIEM LE chuyen khoan ung ho mien trung page_1
13/10/2020 5091.56410 500.000 479627.131020.075913.Ung ho mien trung page_1

3. Export Excel

Export each csv directory to an excel file. Example:

python tools/export-excel.py --csv-dir "data/content/TÀI KHOẢN XXX746 (Pass_ [email protected])/TỪ 13.10.20 ĐẾN 23.11.20/1. TRANG 1 -1000.pdf"
usage: export-excel.py [-h] --csv-dir CSV_DIR [--output-dir OUTPUT_DIR] [--transaction-template TRANSACTION_TEMPLATE] [--filename FILENAME]

optional arguments:
  -h, --help            show this help message and exit
  --csv-dir CSV_DIR     csv dir
  --output-dir OUTPUT_DIR
                        output dir
  --transaction-template TRANSACTION_TEMPLATE
                        dir to save transaction content
  --filename FILENAME   output filename, leave blank to set default

4. Extract dataset

From api responed data, you can extract dataset to train text-recognization model:

 python tools/export-dataset.py 
usage: extract-dataset.py [-h] [--respone-dir RESPONE_DIR] [-a OUTPUT_ANNOTATION] [-i OUTPUT_IMAGE_DIR]

optional arguments:
  -h, --help            show this help message and exit
  --respone-dir RESPONE_DIR
                        dir to api respone
  -a OUTPUT_ANNOTATION, --output-annotation OUTPUT_ANNOTATION
                        path to save annotation file
  -i OUTPUT_IMAGE_DIR, --output-image-dir OUTPUT_IMAGE_DIR
                        path to save annotation file
  • Dataset of first 1000 pages lalebed by google-ocr (~336k): Google Drive
  • Tips: you may want to balance data text type before extract

5. Result

18107 transaction statement pages have been extracted from pdf format: Google Drive - Accuracy >99%.

You might also like...
Go package for OCR (Optical Character Recognition), by using Tesseract C++ library

gosseract OCR Golang OCR package, by using Tesseract C++ library. OCR Server Do you just want OCR server, or see the working example of this package?

Convert PDF/Image to TXT using EasyOcr - the best OCR engine available!
Convert PDF/Image to TXT using EasyOcr - the best OCR engine available!

PDFImage2TXT - DOWNLOAD INSTALLER HERE What can you do with it? Convert scanned PDFs to TXT. Convert scanned Documents to TXT. No coding required!! In

A bot that plays TFT using OCR. Keeps track of bench, board, items, and plays the user defined team comp.
A bot that plays TFT using OCR. Keeps track of bench, board, items, and plays the user defined team comp.

NOTES: To ensure best results, make sure you are running this on a computer that has decent specs. 1920x1080 fullscreen is required in League, game mu

This pyhton script converts a pdf to Image then using tesseract as OCR engine converts Image to Text

Script_Convertir_PDF_IMG_TXT Este script de pyhton convierte un pdf en Imagen luego utilizando tesseract como motor OCR convierte la Imagen a Texto. p

Extract tables from scanned image PDFs using Optical Character Recognition.

ocr-table This project aims to extract tables from scanned image PDFs using Optical Character Recognition. Install Requirements Tesseract OCR sudo apt

Ready-to-use OCR with 80+ supported languages and all popular writing scripts including Latin, Chinese, Arabic, Devanagari, Cyrillic and etc.
Ready-to-use OCR with 80+ supported languages and all popular writing scripts including Latin, Chinese, Arabic, Devanagari, Cyrillic and etc.

EasyOCR Ready-to-use OCR with 80+ languages supported including Chinese, Japanese, Korean and Thai. What's new 1 February 2021 - Version 1.2.3 Add set

A Python wrapper for the tesseract-ocr API

tesserocr A simple, Pillow-friendly, wrapper around the tesseract-ocr API for Optical Character Recognition (OCR). tesserocr integrates directly with

FastOCR is a desktop application for OCR API.

FastOCR FastOCR is a desktop application for OCR API. Installation Arch Linux fastocr-git @ AUR Build from AUR or install with your favorite AUR helpe

OCR-D-compliant page segmentation

ocrd_segment This repository aims to provide a number of OCR-D-compliant processors for layout analysis and evaluation. Installation In your virtual e

Owner
Nguyen Xuan Hung
R&D
Nguyen Xuan Hung
The code for “Oriented RepPoints for Aerail Object Detection”

Oriented RepPoints for Aerial Object Detection The code for the implementation of “Oriented RepPoints”, Under review. (arXiv preprint) Introduction Or

WentongLi 207 Dec 24, 2022
LEARN OPENCV IN 3 HOURS USING PYTHON - INCLUDING EXAMPLE PROJECTS

LEARN OPENCV IN 3 HOURS USING PYTHON - INCLUDING EXAMPLE PROJECTS

Murtaza Hassan 815 Dec 29, 2022
Primary QPDF source code and documentation

QPDF QPDF is a command-line tool and C++ library that performs content-preserving transformations on PDF files. It supports linearization, encryption,

QPDF 2.2k Jan 04, 2023
Pixel art search engine for opengameart

Pixel Art Reverse Image Search for OpenGameArt What does the final search look like? The final search with an example can be found here. It looks like

Eivind Magnus Hvidevold 92 Nov 06, 2022
Opencv-image-filters - A camera to capture videos in real time by placing filters using Python with the help of the Tkinter and OpenCV libraries

Opencv-image-filters - A camera to capture videos in real time by placing filters using Python with the help of the Tkinter and OpenCV libraries

Sergio Díaz Fernández 1 Jan 13, 2022
[EMNLP 2021] Improving and Simplifying Pattern Exploiting Training

ADAPET This repository contains the official code for the paper: "Improving and Simplifying Pattern Exploiting Training". The model improves and simpl

Rakesh R Menon 138 Dec 26, 2022
零样本学习测评基准,中文版

ZeroCLUE 零样本学习测评基准,中文版 零样本学习是AI识别方法之一。 简单来说就是识别从未见过的数据类别,即训练的分类器不仅仅能够识别出训练集中已有的数据类别, 还可以对于来自未见过的类别的数据进行区分。 这是一个很有用的功能,使得计算机能够具有知识迁移的能力,并无需任何训练数据, 很符合现

CLUE benchmark 27 Dec 10, 2022
Image processing using OpenCv

Image processing using OpenCv Write a program that opens the webcam, and the user selects one of the following on the video: ✅ If the user presses the

M.Najafi 4 Feb 18, 2022
Genalog is an open source, cross-platform python package allowing generation of synthetic document images with custom degradations and text alignment capabilities.

Genalog is an open source, cross-platform python package allowing generation of synthetic document images with custom degradations and text alignment capabilities.

Microsoft 235 Dec 22, 2022
Virtual Zoom Gesture using OpenCV

Virtual_Zoom_Gesture I have created a virtual zoom gesture where we can Zoom in and Zoom out any image and even we can move that image anywhere on the

Mudit Sinha 2 Dec 26, 2021
Converts an image into funny, smaller amongus characters

SussyImage Converts an image into funny, smaller amongus characters Demo Mona Lisa | Lona Misa (Made up of AmongUs characters) API I've also added an

Dhravya Shah 14 Aug 18, 2022
Code for paper "Role-based network embedding via structural features reconstruction with degree-regularized constraint"

Role-based network embedding via structural features reconstruction with degree-regularized constraint Train python main.py --dataset brazil-flights

wang zhang 1 Jun 28, 2022
Code release for Hu et al., Learning to Segment Every Thing. in CVPR, 2018.

Learning to Segment Every Thing This repository contains the code for the following paper: R. Hu, P. Dollár, K. He, T. Darrell, R. Girshick, Learning

Ronghang Hu 417 Oct 03, 2022
Responsive Doc. scanner using U^2-Net, Textcleaner and Tesseract

Responsive Doc. scanner using U^2-Net, Textcleaner and Tesseract Toolset U^2-Net is used for background removal Textcleaner is used for image cleaning

3 Jul 13, 2022
Handwriting Recognition System based on a deep Convolutional Recurrent Neural Network architecture

Handwriting Recognition System This repository is the Tensorflow implementation of the Handwriting Recognition System described in Handwriting Recogni

Edgard Chammas 346 Jan 07, 2023
Multi-choice answer sheet correction system using computer vision with opencv & python.

Multi choice answer correction 🔴 5 answer sheet samples with a specific solution for detecting answers and sheet correction. 🔴 By running the soluti

Reza Firouzi 7 Mar 07, 2022
Handwritten Text Recognition (HTR) using TensorFlow 2.x

Handwritten Text Recognition (HTR) system implemented using TensorFlow 2.x and trained on the Bentham/IAM/Rimes/Saint Gall/Washington offline HTR data

Arthur Flôr 160 Dec 21, 2022
An organized collection of tutorials and projects created for aspriring computer vision students.

A repository created with the purpose of teaching students in BME lab 308A- Hanoi University of Science and Technology

Givralnguyen 5 Nov 24, 2021
Forked from argman/EAST for the ICPR MTWI 2018 CHALLENGE

EAST_ICPR: EAST for ICPR MTWI 2018 CHALLENGE Introduction This is a repository forked from argman/EAST for the ICPR MTWI 2018 CHALLENGE. Origin Reposi

Haozheng Li 157 Aug 23, 2022
Memory tests solver with using OpenCV

Human Benchmark project This project is OpenCV based programs which are puzzle solvers for 7 different games for https://humanbenchmark.com/. made as

Bahadır Araz 24 Dec 27, 2022