githubharald/SimpleHTR
A handwritten text recognition system using CNN and RNN with TensorFlow that extracts text from images of words or text lines.
Not currently ranked — collecting fresh signals.
star history
The system uses a convolutional neural network for feature extraction followed by recurrent neural networks (LSTM) for sequence modeling, with CTC (Connectionist Temporal Classification) decoding to handle variable-length output sequences. It is trained on the IAM offline handwritten text recognition dataset and can process single word images or multi-word text lines.
Frequently asked
- What is githubharald/SimpleHTR?
- A handwritten text recognition system using CNN and RNN with TensorFlow that extracts text from images of words or text lines.
- Is SimpleHTR open source?
- Yes — githubharald/SimpleHTR is open source, released under the MIT license.
- What language is SimpleHTR written in?
- githubharald/SimpleHTR is primarily written in Python.
- How popular is SimpleHTR?
- githubharald/SimpleHTR has 2.2k stars on GitHub.
- Where can I find SimpleHTR?
- githubharald/SimpleHTR is on GitHub at https://github.com/githubharald/SimpleHTR.