Authors:

• Stefan Evanghelides - s2895323

• Huaitin Lu - s3339246

• Sudhakaran Jain - s3558487

• Subilal Vattimunda Purayil - s3587630

main.py - the main executable

mainPreprocess.py - executable only for the preprocessing

mainSegment.py - executable only for the segmentation

• Preprocess - module containing the preprocessing steps. Can also be used as a standalone application.

• Segment - module containing the segmentation of an image into lines and, further, segmentation of lines into words. The prerequisite for this step is that the image must be preprocessed.

• Classifier - module containing the classifier used for recognition of the handwriting text. This can also be used as a standalone application.

• Utils - contains common functions for handling reading/writing of data.

(This application is programmed in Python 3. For brevity, python3 commands will be used as python and pip3 command will be used as pip). We have two methods for creating virtual environment (virtualenv or pipenv), choose anyone.

Installation:

• install Python 3 (ideally 3.6.x, but not 3.7.x because it does not support TensorFlow yet)

• install virtualenv using pip install virtualenv
• navigate to the main folder
• create virtual environment using virtualenv -p python3 env
• activate the environment using source env/bin/activate
• install all the required libraries using pip install -r requirements.txt

• install pipenv using pip install pipenv
• install the program using pipenv install when you are in the main folder.

Notes about pipenv:

• If pipenv is used (or perhaps virtualenv), then pipenv run must be appended to the following commands.

• python main.py input_folder

For preprocessing:

• python mainPreprocess.py -> the input file must be manually typed in.

For segmentation:

• python mainSegment.py -> the input file must be manually typed in.

For Classifier:

• python Classifier\train_network.py -> for training the network

• python mainClassifier.py -> for classification. Input file must be manually typed in.

1. Command line help:

   when using pipenv: pipenv run python main.py -h

   when using virtualenv: python main.py -h

   Output:

python main.py -h
usage: main.py [-h] [-d | -f] [-e EXTENSION] input_dir

positional arguments:
  input_dir             Specify the input directory

optional arguments:
  -h, --help            show this help message and exit
  -d, --debug           run program in debug mode
  -f, --fast            skip over intermediary results (used to speed-up the
                        program)
  -e EXTENSION, --extension EXTENSION
                        specify the extension (default=.jpg)  
  -o OUTPUT, --output OUTPUT
                        specify the output directory

2. Sample running 4 images of about 2156 x 1625 in FAST mode:

   Using the command:
   pipenv run python main.py -f InputTest -o OutputTest -- incase of pipenv OR,
   python main.py -f InputTest -o OutputTest -- incase of virtualenv

Output:

pipenv run python main.py -f InputTest -o OutputTest

Processing files from directory:   InputTest
With extension:                    *jpg
Output Directory:                  OutputTest
Runmode:                           NORMAL


Processing image1.jpg
Finished preprocessing image1.jpg
    ****
Segmenting image1.jpg
Finished segmenting image1.jpg
    ****
Classifying image1.jpg
Finished classifying image1.jpg

Output is in: THIS-IS-PRIVATE-PATH\OutputTest\image1\output.txt

Elapsed time: 10.896777629852295
-----------------------------------

Processing image2.jpg
Finished preprocessing image2.jpg
    ****
Segmenting image2.jpg
Finished segmenting image2.jpg
    ****
Classifying image2.jpg
Finished classifying image2.jpg

Output is in: THIS-IS-PRIVATE-PATH\OutputTest\image2\output.txt

Elapsed time: 11.643365621566772
-----------------------------------

Processing image3.jpg
Finished preprocessing image3.jpg
    ****
Segmenting image3.jpg
Finished segmenting image3.jpg
    ****
Classifying image3.jpg
Finished classifying image3.jpg

Output is in: THIS-IS-PRIVATE-PATH\OutputTest\image3\output.txt

Elapsed time: 11.344895839691162
-----------------------------------

Processing image4.jpg
Finished preprocessing image4.jpg
    ****
Segmenting image4.jpg
Finished segmenting image4.jpg
    ****
Classifying image4.jpg
Finished classifying image4.jpg

Output is in: THIS-IS-PRIVATE-PATH\OutputTest\image4\output.txt

Elapsed time: 10.71731948852539
-----------------------------------
Total elapsed time: 44.60835003852844
Average/image = 11.150589644908905
-----------------------------------