Robust ReadingCompetition

method: TencentAILab2018-04-24

Authors: Jingchao Zhou, Tianlin Gao, Zheng Zhou, Zhifeng Li

Description: We train a network to recognize the word images. First, we correct the oblique and vertical arranged text lines using tranditional OCR technologies. Second, we generate several batches of synthesized images with similar style and arrangement as training samples. Last, we adopt DenseNet as the backbone to extract features, Bi-direction LSTM to learn sequential information, and CTC as the transcription layer.

method: Tencent-OCR+2017-06-30

Authors: Chunchao Guo, Weichen Zhang, Yi Li, Hui Song, Ming Liu, Hongfa Wang, Lei Xiao

Description: Data Platform Department, Tencent. We adapt CNN-LSTM-CTC architecture to recognize the text line. In addition, a knowledge-based post processing is used for adjusting the result.

method: HIK_OCR2017-07-01

Authors: Zhanzhan Cheng*, Gang Zheng*, Fan Bai, Yunlu Xu, Jie Wang, Yangliu Xu, Ying Yao, Fan Wu, Yi Niu(*equal contribution)

Description: Method is designed based on the sequence-sequence framework:
Encoder: Images are resized to 100pixels x 100pixels, and features are extracted by using complicated CNN;
Decoder: Character sequence generation with Attention-based decoder.

1)We design a complicated CNN-based feature extraction mechanism(Mask spatial transform etc.) for capturing arbitrary placed text features;
2)In order to handle the character additions or deletions problem, we develop an Edit Probalibilty Loss instead of the SofmaxWithLoss in the sequence learning task.
3)We also design a self-adaption gate mechanism for CNN so that network can capture global information.

The papers are in preparation.

Ranking Table

Description Paper Source Code
DateMethodTotal Edit distance (case sensitive)Correctly Recognised Words (case sensitive)T.E.D. (case insensitive)C.R.W. (case insensitive)

Ranking Graphic

Ranking Graphic