VietnameseOCR - Vietnamese Optical Character Recognition Apply Deep Learning ( CNN networks ) to train a model uses for recognizing Vietnamese characters, it works well with Latin characters.
Dataset in big image ( 10.000 samples, 2800 x 2800 pixel)
python 3.6.5tensorflowPILLayer Shape Kernel Stride Padding INPUT [28, 28, 1] CONV1 [3, 3, 32, 32] [1, 1] SAME POOL1 CONV2 [3, 3, 32, 64] [1, 1] SAME POOL2 CONV3 [3, 3, 64, 128] [1, 1] SAME POOL3 FC1 FC2 [625, 190]
Training.........Epoch: 38 cost = 0.312853018Epoch: 39 cost = 0.298816641Epoch: 40 cost = 0.293328794Evaluation------------------------------Test Accuracy: 0.974867469544Prepare dataset for training git clone https://github.com/miendinh/VietnameseOCR.gitcd VietnameseOCR/data/train/charactersunzip dataset.zipPrepare fonts for generating text-image cd VietnameseOCR/data/train/charactersunzip google.zipunzip win.zipCreate font list, then save it in fonts.list Generate Text Image Dataset Play with pretrained model All pretrained weights of model is save to file vocr.brain Let's test with random character in dataset STN-OCR: A single Neural Network for Text Detection and Text Recognition Automatic Dataset Augmentation VGG16 implementation in TensorFlow Vietnamese Dict (VietOCR3) Author mien.hust [at] gmail [dot] com 
