crnn_ctc

CRNN_CTC_PyTorch

This version is an improvement based on Mei used to train CRNN model on PyTorch1.0. Relevant paper is "An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition" by Shi et. al. The filter.py is completed by Xu to exclude the influence by the corrupted images.

To run a demo, please refer to Mei.

Train a model

To train a model (I used SynthText90k dataset to train), check the details in the code and run like the following steps:

1. Run filter.py for file annotation_train.txt and annotation_val.txt.
2. Run python saveAsLmdb.py.
3. Run python crnn_main.py --cuda.