---
owner:
  firstname: "Limeng"
  lastname: "Liu"
  hid: "fa19-523-202"
  community: "423"
  semester: "fa19"


# FA19 423 Final Project

* Using the IAM off-line Handwritten Text Recognition dataset for implementing a deep learning project.
* Neural Network created by TensorFlow (TF) and trained on the IAM dataset
* I used the words images for testing and used the sentences for training
* 3/4 of the words from the validation-set are correctly recognized and the character error rate is around 10%

## Folder/File Explanation

* i423_project_fa19-523-202_Limeng_Liu.ipynb: Main project file
* i423_project_fa19-523-202_Limeng_Liu.pdf: Final Paper
* Image Processing Testing.ipynb: Image processing for sentences samples
* testing code.ipynb: Prelim try for projects with testing codes
* forms_for_parsing.txt: Image forms for parsing
* checkpoint2: Model epoches checkpoints
* data_subset: dataset of 50 writters handwriting-database of text-line sentence
* data-words: IAM dataset of words images and forms for parsing
* Report(LaTex).zip: Latex file for the final report


## References

\[1\] [The IAM-database: an English sentence database for offline handwriting recognition](http://www.fki.inf.unibe.ch/databases/iam-handwriting-database)

\[2\] [Handwritten Text Recognition (OCR) with MXNet Gluon] (https://github.com/awslabs/handwritten-text-recognition-for-apache-mxnet)

\[3\] [Handwritten Text Recognition with TensorFlow] (https://github.com/githubharald/SimpleHTR)

\[4\] [IAM Handwriting Top50: Offline IAM Handwriting Dataset's subset, w.r.t. the 50 most common writers.] (https://www.kaggle.com/tejasreddy/iam-handwriting-top50)

\[5\] [Build a Handwritten Text Recognition System using TensorFlow: A minimalistic neural network implementation which can be trained on the CPU] (https://towardsdatascience.com/build-a-handwritten-text-recognition-system-using-tensorflow-2326a3487cd5)

\[6\] [English Deep Writer] (https://github.com/priya-dwivedi/Deep-Learning/blob/master/handwriting_recognition/English_Writer_Identification.ipynb)

\[7\] [DeepWriter: A Multi-Stream Deep CNN for Text-independent Writer Identification] (https://arxiv.org/pdf/1606.06472.pdf)

\[8\] [Handwriting recognition using Tensorflow and Keras] (https://towardsdatascience.com/handwriting-recognition-using-tensorflow-and-keras-819b36148fe5)

\[9\] [Keras CNN training to recognize captcha: Get low loss and get low accuracy] (https://stackoverflow.com/questions/53993955/keras-cnn-training-to-recognize-captcha-get-low-loss-and-get-low-accuracy)