Braille is an important communication code for blind people. The automatic recognition of braille is adapted to convex braille points. This paper describes the technology for recognizing mixed convex and concave braille points that result from typing on both sides of a sheet of paper. The braille points are recognized by scanning only one side and are translated into the kana syllabary. Convex points can be distinguished from concave points by the highlights and shadow patterns generated by exposing the braille points to LED light. The braille points which form one character are extracted by applying an extraction mask. The recognition performance of this method is evaluated by using the prototype system in the presence of deformed braille points, deteriorated paper, and slanting paper. The recognition ratio is estimated to be more than 96 percent for five different braille patterns prepared by various typing methods, including a braille printer.
机构:
Oklahoma State Univ, Sch Elect & Comp Engn, Stillwater, OK 74078 USA
Oklahoma State Univ, Ctr Vet Hlth Sci, Dept Vet Clin Sci, Stillwater, OK 74078 USAOklahoma State Univ, Sch Elect & Comp Engn, Stillwater, OK 74078 USA
Piao, Daqing
论文数: 引用数:
h-index:
机构:
Fixler, Dror
NANOSCALE IMAGING, SENSING, AND ACTUATION FOR BIOMEDICAL APPLICATIONS XVI,
2019,
10891