NOT COMPLETED PROJECT for OCR

In this repo there is an attempt to create an OCR model for the first time

├── Interface
│   ├── google_app
│   ├── interface
├── data
├── notebooks
├── src
│   ├── dataset
│   ├── text encoding
│   ├── text processing
│   ├── training
│   └── inference
└── utils

The repository is organized as follows:

• Interface/: This directory is the django project and contains google_app as the django app.

• data/: This directory contains the dataset used for training and evaluation. It includes both the raw data and preprocessed versions, if applicable.

• notebooks/: This directory contains Jupyter notebooks that provide step-by-step explanations of the data exploration, preprocessing, model training, and evaluation processes. Explore if you are feeling adventurous

• src/: This directory contains the source code for the project, including data preprocessing scripts, model training scripts, and evaluation scripts.

• utils/: This directory contains utility functions and helper scripts used throughout the project.

• It is recommended to set up a virtual environment for this project using python 3.8.16

To get started with the project, follow these steps:

1. Clone the repository:

   git clone https://github.com/Aylore/blnk-OCR.git
   

2. Change directory into the repository:

   cd blnk-OCR
   

3. Install the required dependencies:

   make install
   

1. Text Processing

   I tired various methods for text encoding , none of them seemed to get good results so I just setteled for simple encoding

2. Image Processing

   The images were resized to 50 * 200 I extracted the text only in new images with all black background but seemed to confuse the model for some reason

3. Training

   A CRNN (CNN + RNN ) model for used for seuquence generation

4. Django Integration

   After training I used django to create and easy to use not too fancy GUI.

1. Use

 make run

If you would like to contribute to this project, Feel Free to make a pull request

• Extract characters one by one using CV and start classifying them by a pretrained model
• Use attention based model