vmd7 / automate-identification-and-recognition-of-handwritten-text-from-an-image

This is simple and easy way to identification and recognition of handwritten text from an image. An Optical Character Recognition (OCR) is type of image-based sequence recognition problem. So, i used CRNN to cop up this problem. This will help to extract text from an image.

Automate Identification and Recognition of Handwritten Text from an Image

In this repository i used various techniques to identify and recognize the text from an image.

In Second Folder "HTR_Using_CRNN" i used CRNN concept to identify and recognize the text from an image.

1) Handwritten Text Recognition Using CRNN

To use this repository on your loacal machine or Google Colab follow the below steps:

For Google Colab

1. First Clone and Download this repository.

2. Then download the handwritten IAM dataset from this link - (http://www.fki.inf.unibe.ch/databases/iam-handwriting-database/download-the-iam-handwriting-database) or (http://www.fki.inf.unibe.ch/DBs/iamDB/data/words)

3. Creat your kaggle account and upload this data in kaggle.

4. Go to "Your Account" and create New "API" and download the .json file.

5. Now upload the "words.txt" to your Google Drive which is in "Data" folder of "HTR_Using_CRNN" folder.

6. Change necessary directories such as words.txt, .json file locations in code.

7. Now follow the HTR_Using_CRNN.ipynb

For Local Machine

Skip steps 2,3,4,5 also don't use Kaggle setting up section code in "HTR_Using_CRNN.ipynb" file.