Amharic Character Image Recognition

In this paper we introduce Convolutional Neural Network (CNN) based method for Amharic character image recognition. We also introduce a dataset for training purposes. The proposed method has less pre-processing steps and out per- forms the state-of-the-art by a large margin. Experiments were done on 80,000 Amharic character images which was generated with different degradation level. We systematically evaluated the performance of the recognition model and achieved the state-of- art performance with an average recognition accuracy of 92.71%.

[1]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[2]  Debashis Ghosh,et al.  Script Recognition—A Review , 2010, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[3]  Karl Tombre,et al.  The Evolution of Document Image Analysis , 2014, Handbook of Document Image Processing and Recognition.

[4]  Bo Xu,et al.  Image character recognition using deep convolutional neural network learned from different languages , 2014, 2014 IEEE International Conference on Image Processing (ICIP).

[5]  Thomas M. Breuel,et al.  High-Performance OCR for Printed English and Fraktur Using LSTM Networks , 2013, 2013 12th International Conference on Document Analysis and Recognition.

[6]  Lianwen Jin,et al.  Recognition confidence analysis of handwritten Chinese character with CNN , 2015, 2015 13th International Conference on Document Analysis and Recognition (ICDAR).

[7]  Malayappan Shridhar,et al.  Offline Handwritten Devanagari Word Recognition: A Holistic Approach Based on Directional Chain Code Feature and HMM , 2008, 2008 International Conference on Information Technology.

[8]  Chokri Ben Amar,et al.  Dyadic Multi-resolution Analysis-Based Deep Learning for Arabic Handwritten Character Classification , 2015, 2015 IEEE 27th International Conference on Tools with Artificial Intelligence (ICTAI).

[9]  Sherief Abdallah,et al.  Improving Handwritten Arabic Character Recognition by Modeling Human Handwriting Distortions , 2015, ACM Trans. Asian Low Resour. Lang. Inf. Process..

[10]  Yajie Liu,et al.  Offline handwritten English character recognition based on convolutional neural network , 2012, 2012 10th IAPR International Workshop on Document Analysis Systems.

[11]  Lianwen Jin,et al.  High performance offline handwritten Chinese character recognition using GoogLeNet and directional feature maps , 2015, 2015 13th International Conference on Document Analysis and Recognition (ICDAR).

[12]  Thomas M. Breuel,et al.  The OCRopus open source OCR system , 2008, Electronic Imaging.

[13]  Lianwen Jin,et al.  Improved deep convolutional neural network for online handwritten Chinese character recognition using domain-specific knowledge , 2015, 2015 13th International Conference on Document Analysis and Recognition (ICDAR).

[14]  Yong Hu,et al.  Character Recognition via a Compact Convolutional Neural Network , 2017, 2017 International Conference on Digital Image Computing: Techniques and Applications (DICTA).

[15]  Abdul Kawsar Tushar,et al.  Handwritten Arabic numeral recognition using deep learning neural networks , 2017, 2017 IEEE International Conference on Imaging, Vision & Pattern Recognition (icIVPR).

[16]  Yaregal Assabie.  Optical Character Recognition of Amharic Text: an Integrated Approach , 2002 .

[17]  K. P. Soman,et al.  Convolutional Neural Networks for the Recognition of Malayalam Characters , 2014, FICTA.