An effective recognition method for medical sheet based on deep learning approach

At present, most medical sheet (such as medical report, laboratory sheet, medical cases, etc.) in the form of nonelectronic is easy to lose, and difficult to be integrated with other electronic health data. In order to fully utilize these valuable data, in this paper we propose a deep learning approach, named k-CNN, which can intelligently recognize the contents of medical sheet. The main advantages of k-CNN are summarized as follows: Firstly, the local feature of the character is extracted by the pattern recognition KNN algorithm; Secondly, CNN algorithm is used to extract the deep features of the characters; Finally, based on the tesseract that is an open source recognition engine, the recognition results of KNN and CNN are combined to get the intelligent identification of medical sheet. The experimental results show that k-CNN recognition algorithm can accurately identify the common medical sheets, and the recognition accuracy is better than other traditional algorithms.

[1]  Rita Cucchiara,et al.  Layout analysis and content enrichment of digitized books , 2014, Multimedia Tools and Applications.

[2]  David S. Doermann,et al.  Identifying script on word-level with informational confidence , 2005, Eighth International Conference on Document Analysis and Recognition (ICDAR'05).

[3]  Miguel Angel Ferrer-Ballester,et al.  LBP Based Line-Wise Script Identification , 2013, 2013 12th International Conference on Document Analysis and Recognition.

[4]  Jiwen Lu,et al.  Deep hashing for compact binary codes learning , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[5]  Xuelong Li,et al.  Principal Component 2-D Long Short-Term Memory for Font Recognition on Single Chinese Characters , 2016, IEEE Transactions on Cybernetics.

[6]  Yueting Zhuang,et al.  Latent Style Model: Discovering writing styles for calligraphy works , 2009, J. Vis. Commun. Image Represent..

[7]  Rémy Mullot,et al.  Texture feature benchmarking and evaluation for historical document image analysis , 2017, International Journal on Document Analysis and Recognition (IJDAR).

[8]  Tieniu Tan,et al.  Font Recognition Based on Global Texture Analysis , 2001, IEEE Trans. Pattern Anal. Mach. Intell..

[9]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[10]  Pengcheng Gao,et al.  Chinese calligraphic style representation for recognition , 2017, International Journal on Document Analysis and Recognition (IJDAR).