Vision-based patient identification recognition based on image content analysis and support vector machine for medical information system

• DOI: 10.1186/s13634-020-00686-3
• Publication Date: 2020-05-29T09:02:43Z
• Authors: Guo-Shiang Lin; Sin-Kuo Chai; Hsiang-Min Li; Jen-Yung Lin
• Keywords: Optical character recognition; Normalization; Numerical digit

Abstract In this paper, a vision-based patient identification recognition system based on image content analysis and support vector machine is proposed for medical information system, especially in dermatology. This composed of three parts: pre-processing, candidate region detection, digit recognition. To consider the efficiency scheme, normalization performed. The color used to identify camera-captured screen images. pre-processing part, effect noise captured images reduced by bilateral filter. spatial initially roughly locate region. reduce skew effect, correction algorithm Hough transform developed. A template matching find special symbols locating interest (ROI). For segmentation, digits are segmented ROI vertical projection adaptive thresholding. recognition, some features measured from each segment classifier applied recognize digits. experiment’s results show that could effectively not only use distinguish skin but also detect ROIs. After accuracy rates 98.4% 94.2% Tesseract Optical Character Recognition (OCR) software, respectively. These demonstrate outperforms OCR software terms rate