tesseract-ocr-re's Introduction

Tesseract 4 OCR Runtime Environment - Docker Container

Based on Alexander Pozdnyakov's Personal Package Archive tesseract-ocr - Tesseract command line OCR tool. If you want to compile Tesseract 4 yourself please take look at tesseractshadow/tesseract4cmp.

If you are not familiar with Docker please read Docker - Get Started.

Quick start

Prerequisites:

Install Docker
Download and unzip this repository
Pull tesseractshadow/tesseract4re image from Docker Hub (automated build using Dockerfile from this repository)
```
docker pull tesseractshadow/tesseract4re
```

Scripted steps (cd ./scripts/):

Run the image as the t4re ./run.sh or just start it if is stopped ./start.sh.
Do some OCR test: ./test.sh.
Stop t4re container instance: ./stop.sh.

tesseract-ocr-re's People

Contributors

Stargazers

Watchers

tesseract-ocr-re's Issues

Arabic traineddata causes segfault

Using ara.traineddate for version 4.0 of tessaract, obtained here, causes the following error when running test.sh:

====== TASK TASK_12802_394560981 started ======
mgr->GetComponent(TESSDATA_INTTEMP, &fp):Error:Assert failed:in file adaptmatch.cpp, line 537
/bin/bash: line 1:    81 Segmentation fault      (core dumped) tesseract ../myfile.jpg myfile -l ara --psm 1 --oem 2 txt pdf hocr
====== Result files was copied to ./ocr-files/output/TASK_12802_394560981/ ======

What is the cause of this error?
And how can I resolve this?

changes in apt label

FROM tesseractshadow/tesseract4re
RUN apt-get update

E: Repository 'http://ppa.launchpad.net/alex-p/tesseract-ocr/ubuntu bionic InRelease' changed its 'Label' value from 'tesseract-ocr - Tesseract command line OCR tool' to 'tesseract-ocr - Tesseract command line OCR tool (stable)'

Looping through images

Firstly, the program is very good and easily downloadable, installable, and operable.

Two things I wish to have that might not be difficult for the author to provide.

Commenting code esp in the Dockerfile and script files.
Suggest how one can loop through images to get tesseract to recognize multiple images and store the outputs (hOCR, PDF, and text) of each image in a different output file.

Appreciate if an early response may ensue. BTW, I have found a workaround but it may not be the best one.

Thanks

Recommend Projects