Handwritten Text Recognition (HTR) system implemented with TensorFlow (TF) and trained on the IAM off-line HTR dataset. The model takes images of single words or text lines (multiple words) as input and outputs the recognized text. 3/4 of the words from the validation-set are correctly recognized, and the character error rate is around 10%.

Run demo

  • Download one of the pretrained models

    • Model trained on word images: only handles single words per image, but gives better results on the IAM word dataset https://www.dropbox.com/s/mya8hw6jyzqm0a3/word-model.zip?dl=1

    • Model trained on text line images: can handle multiple words in one image https://www.dropbox.com/s/7xwkcilho10rthn/line-model.zip?dl=1

  • Put the contents of the downloaded zip-file into the model directory of the repository
  • Go to the src directory
  • Go to main.py and install required packages numpy, tensorflow, editpath, OpenCV using pip install
  • The input images, and the expected outputs are shown below when the text line model is used.

image

Init with stored values from ../model/snapshot-13
Recognized: “or work on line level”
Probability: 0.6674368977546692

image

Init with stored values from ../model/snapshot-13
Recognized: “word”
Probability: 0.9806373119354248

Unable to Recognize Every Image

image

Init with stored values from ../model/snapshot-13
Recognized: “fnd”
Probability: 0.026317203417420387

But when resized my image according to that sample image it was nearly able to recognize

wordtest

Init with stored values from ../model/snapshot-13
Recognized: “wword”
Probability: 0.6883667707443237

<
Previous Post
What is Git?
>
Next Post
What is DAO?